TROPONIN TESTING FOR DIAGNOSIS OF ACUTE CORONARY … · ACS Acute coronary syndrome AF ... CABG Coronary artery bypass graft surgery CASP Critical Appraisal Skills Programme CHD Coronary

1

TROPONIN TESTING FOR DIAGNOSIS OF ACUTE CORONARY SYNDROMES IN WESTERN AUSTRALIAN URBAN PRIMARY CARE

Helen Mary Wilcox, MB BS(Hons) FRACGP DCH

This thesis is presented for partial fulfilment of the degree of Master of Clinical

Research of The University of Western Australia

School of Primary, Aboriginal and Rural Health Care

Faculty of Medicine, Dentistry and Health Sciences

University of Western Australia

December 2015

2

ABSTRACT Objective: To examine primary care use of cardiac troponin (cTn) testing for acute

coronary syndrome (ACS) diagnosis.

Design: Prospective cohort study.

Setting: General practitioner-initiated cTn tests conducted from 24 September 2009 to

3 September 2010 in Perth, Western Australia.

Participants: 369 patients with samples collected at laboratory community collection

centres. Requesting GPs provided the clinical context for testing (124 GPs, response

rate 38%). Patient outcomes were obtained from linked data sources for 12 months

following the final test. Clinical information and outcomes were compared with data

from emergency department patients with ACS symptoms.

Main outcome measures: Cardiovascular risk status, symptoms prompting cTn

testing; estimated ACS likelihood and referral decision before and after testing; result

turnaround time; hospital presentations, procedures and mortality.

Results: Of the 328 GPs who received a survey request, 124 (37.8%) responded.

122/124 (98.4%) of test results were negative. 71/104 patients (69%) were at high or

intermediate risk of ACS based on clinical risk factors. 69/124 patients (55.6%) had

typical ischaemic pain and 62/124 patients (50.0%) were tested within 48 h of symptom

onset (23.4% within 12 h, with no serial testing). Test results affected GPs’ estimation

of ACS likelihood (P < 0.01) but not referral decisions (P = 0.23). 94/355 patients

(26.5%) presented to hospital with cardiovascular symptoms or diagnoses during

follow-up; 27/355 patients (7.6%) had at least one ACS, 13/255 (3.7%) within 1 month

of testing.

Conclusions: GP-initiated cTn testing involves patients at high risk of ACS. ACS and

associated adverse outcomes can occur in patients undergoing testing, even when the

cTn test result is negative. Potential gaps exist in physicians’ understanding of the

limitations of cTn testing, and results have minimal influence on their patient

management. GPs may benefit from guidance in ordering cTn testing.

3

TABLE OF CONTENTS

ABSTRACT .................................................................................................................. 2

TABLE OF CONTENTS ............................................................................................... 3

LIST OF TABLES ......................................................................................................... 6

LIST OF FIGURES ....................................................................................................... 7

GLOSSARY ................................................................................................................. 8

CHAPTER ONE: BACKGROUND.............................................................................. 11

1. ACUTE CORONARY SYNDROMES IN AUSTRALIA ........................................ 11 1.1 Acute coronary syndromes: prevalence and burden ............................ 11 1.2 ACS: pathogenesis, classification and diagnosis .................................. 11

2. TROPONIN ....................................................................................................... 12 2.1 The troponin complex........................................................................... 12 2.2 Troponin in diagnosis and risk stratification of ACS .............................. 13 2.3 Troponin in non-ACS contexts ............................................................. 15 2.4 Delay in troponin testing ....................................................................... 16 2.5 Monitoring during troponin testing ........................................................ 17 2.6 Highly sensitive troponin in ACS diagnosis .......................................... 17

3. TROPONIN AND ACS IN AUSTRALIAN GENERAL PRACTICE ....................... 18 3.1 Clinical diagnosis of ACS in general practice ....................................... 18 3.2 Benefits and limitations of troponin testing in general practice ............. 19 3.3 Practical issues in troponin testing in primary care ............................... 22 3.4 Estimates of frequency of troponin testing in primary care ................... 22

4. GUIDELINES FOR PRIMARY CARE USE OF TROPONIN ............................... 23 4.1 Australian guidelines ............................................................................ 23 4.2 International guidelines ........................................................................ 23

5. CONCLUSION ................................................................................................... 24 AIMS .......................................................................................................................... 25

HYPOTHESES ........................................................................................................... 25

CHAPTER TWO: SYSTEMATIC REVIEW ................................................................. 26

1. RATIONALE ...................................................................................................... 26 2. OBJECTIVES .................................................................................................... 26 3. METHODS ........................................................................................................ 26

3.1 Theoretical and methodological approach ............................................ 26 3.2 Eligibility criteria ................................................................................... 26 3.3 Information sources ............................................................................. 28 3.4 Search ................................................................................................. 28 3.5 Study selection..................................................................................... 30 3.6 Data collection process ........................................................................ 30 3.7 Data items ............................................................................................ 30 3.8 Risk of bias in individual studies ........................................................... 30 3.9 Synthesis of results .............................................................................. 31

4. RESULTS .......................................................................................................... 32 4.1 Study selection..................................................................................... 32 4.2 Objective 1 ........................................................................................... 36 4.3 Objective 2. .......................................................................................... 39

5. DISCUSSION .................................................................................................... 42 5.1 Summary of evidence .......................................................................... 42 5.2 Limitations ............................................................................................ 45 5.3 Implications for future research ............................................................ 48

6. CONCLUSION ................................................................................................... 49 CHAPTER THREE: METHODS ................................................................................ 50

4

1. STUDY DESIGN ................................................................................................ 50 2. SETTING ........................................................................................................... 50 3. PARTICIPANTS ................................................................................................. 50

3.1 GP cohort ............................................................................................. 50 3.2 ED cohort ............................................................................................. 50 3.3 Linked data .......................................................................................... 51

4. VARIABLES ....................................................................................................... 52 4.1 GP cohort ............................................................................................. 52 4.2 ED cohort ............................................................................................. 53 4.3 Linked data .......................................................................................... 53 4.4 Outcomes ............................................................................................. 54

5. MEASUREMENT ............................................................................................... 56 5.1 GP cohort ............................................................................................. 56 5.2 ED cohort ............................................................................................. 57 5.3 Linked data .......................................................................................... 57

6. STUDY SIZE...................................................................................................... 58 7. QUANTITATIVE VARIABLES ............................................................................ 58 8. STATISTICAL METHODS ................................................................................. 59 CHAPTER FOUR: RESULTS ..................................................................................... 60

1. PARTICIPANTS ................................................................................................. 60 2. DESCRIPTIVE DATA ........................................................................................ 61

2.1 Characteristics of GP cohort with survey data ...................................... 61 2.2 Comparison with Emergency Department cohort ................................. 61 2.3 Patients presenting within 12 hours of symptom onset ......................... 62 2.4 Test result availability ........................................................................... 64 2.5 Effect of test result on GPs’ estimation of likelihood of ACS ................. 65 2.6 Effect of test result on GPs’ intended management .............................. 65

3. OUTCOME DATA .............................................................................................. 66 3.1 Emergency Department cardiovascular presentations.......................... 66 3.2 Admissions and procedures ................................................................. 67 3.3 Time to first hospital presentation ......................................................... 69 3.4 Adverse events .................................................................................... 70 3.4 Outcomes for survey patients in GP cohort .......................................... 71

CHAPTER FIVE: DISCUSSION ................................................................................. 72

1. KEY FINDINGS ................................................................................................. 72 1.1 Clinical characteristics of GP cohort ..................................................... 72 1.2 GPs’ knowledge of troponin’s use and limitations ................................. 72 1.3 Effect of troponin on estimation of ACS likelihood and management .... 72 1.4 Outcomes of patients who underwent troponin testing in primary care . 73

2. INTERPRETATION............................................................................................ 74 2.1 Clinical characteristics of GP cohort ..................................................... 74 2.2 GPs’ knowledge of troponin’s use and limitations ................................. 74 2.3 Effect of troponin on estimation of ACS likelihood and management .... 75 2.4 Outcomes of patients who underwent troponin testing in primary care . 75

3. STRENGTHS AND LIMITATIONS ..................................................................... 78 3.1 Strengths .............................................................................................. 78 3.2 Limitations ............................................................................................ 78

4. GENERALISABILITY AND FUTURE IMPLICATIONS ....................................... 81 4.1 Generalisability of findings ................................................................... 81 4.2 Future use of troponin testing in primary care ...................................... 81 4.3 Future implications for primary care research ....................................... 82

5. CONCLUSION ................................................................................................... 84

5

REFERENCES ........................................................................................................... 85

APPENDICES ............................................................................................................ 97

Appendix 1: Search strategy ........................................................................... 97 Appendix 2: Additional literature sources....................................................... 100 Appendix 3: GP cohort questionnaire ............................................................ 101

6

LIST OF TABLES Table 1: Definition of myocardial infarction .................................................................. 14 Table 2: Non-coronary causes of elevated troponin .................................................... 15 Table 3: Eligibility criteria for systematic review .......................................................... 27 Table 4: List of terms used in electronic search strategy ............................................. 29 Table 5: List of included studies .................................................................................. 33 Table 6: Assessment of risk of bias in included studies using CASP tools .................. 35 Table 7: Results of studies describing outcomes......................................................... 37 Table 8: Results of studies describing GP understanding, use of troponin .................. 40 Table 9: Data variables obtained by laboratories ........................................................ 52 Table 10: Data variables obtained by laboratories from requesting GP ....................... 52 Table 11: Data variables obtained from the ED cohort dataset.................................... 53 Table 12: Data variables from Data Linkage System ................................................... 53 Table 13: EDDC cardiovascular presenting problems ................................................. 54 Table 14: EDDC cardiovascular presenting problems - ICD-10AM codes ................... 55 Table 15: EDDC and HMDC diagnoses ...................................................................... 55 Table 16: EDDC and HMDC diagnoses - ICD-10AM codes ........................................ 56 Table 17: ACHI codes for procedures ......................................................................... 56 Table 18: Adverse outcomes and ICD-10AM codes .................................................... 56 Table 19: Grouping of quantitative variables ............................................................... 58 Table 20: Characteristics of included patients ............................................................. 62 Table 21: Comparison of risk factors between GP and ED cohorts ............................. 62 Table 22: Characteristics of GP patients presenting within 12 hours ........................... 63 Table 23: Time in minutes from specimen collection to result availability .................... 64 Table 24: Effect of test result on GPs' estimation of likelihood of ACS ........................ 65 Table 25: Effect of test result on GPs' intended management ..................................... 65 Table 26: ED presentations - GP cohort ...................................................................... 66 Table 27: ED CVS diagnoses - GP cohort .................................................................. 67 Table 28: Admissions with CVS diagnoses - GP cohort .............................................. 68 Table 29: Details of admission CVS diagnoses - GP cohort ........................................ 68 Table 30: Procedures performed on admitted patients - GP cohort ............................. 68 Table 31: Adverse events within 30 days - GP and ED cohorts .................................. 70

7

LIST OF FIGURES

Figure 1: Classification of acute coronary syndromes ................................................. 12 Figure 2: The troponin complex. ................................................................................. 13 Figure 3: Effect of GP-initiated troponin test results on ACS management ................. 19 Figure 4: Flow diagram of study selection process. ..................................................... 32 Figure 5: Flow chart of participants ............................................................................. 60 Figure 6: Number of patients with typical pain, symptoms <12h and CHD risk factors. 63

8

GLOSSARY

AAA Abdominal aortic aneurysm ACS Acute coronary syndrome AF Atrial fibrillation AO Adverse outcome AV Atroventricular CABG Coronary artery bypass graft surgery CASP Critical Appraisal Skills Programme CHD Coronary heart disease CK Creatine kinase Cr Creatinine CVS Cardiovascular Dx Diagnosis ECG Electrocardiogram ED Emergency department EDDC Emergency Department Data Collection eGFR estimated glomerular filtration rate GP General practitioner HMDC Hospital Morbidity Data Collection hsTn high sensitivity troponin ICU Intensive care unit IQR Interquartile ratio LBBB left bundle branch block MI Myocardial infarction MRDC Mortality Record Data Collection NA Not applicable NHMRC National Health and Medical Research Council NR Not reported NSTEMI Non-ST elevation myocardial infarction PCI Percutaneous coronary intervention POCT Point of care testing RACGP Royal Australian College of General Practitioners RCT randomised controlled trial SJOG St John of God STEMI ST elevation myocardial infarction SVT supraventricular tachycardia TIMI Thrombolysis in Myocardial Infarction TnI Troponin I TnT Troponin T UA Unstable angina URL upper reference limit VT ventricular tachycardia WADLS Western Australian Data Linkage service +ve Positive -ve negative

9

ACKNOWLEDGEMENTS

I would like to acknowledge the expertise, consideration and enthusiasm of my

two supervisors, Professor Jon Emery and Professor Alistair Vickery. Your insight

and wisdom have been invaluable and will shape my research and writing in the

future.

I am grateful to Professor John Burnett and Dr Glenn Edwards from the

supporting laboratories, who facilitated the provision of data and research

assistants, and also gave useful advice for the methodology of this project.

Amanda Hooper and Jenny McMahon, the laboratory based research assistants,

were always conscientious and professional, and provided impetus for the

completion of data collection.

Yusuf Nagree and Stephen MacDonald kindly provided the MIMiC dataset and

were instrumental in obtaining timely relevant HREC approvals.

David Whyatt, Matthew Tuson and Noreen Kirkman generously provided

statistics and library help of their own volition.

The initial development of the research proposal for this thesis was funded by a

Primary Health Care Research, Evaluation and Development (PHCRED)

Research Scholarship.

The contribution of my parents, both academic writers in their own branches of

science and medicine, is very much appreciated. Thank you for the chapter

reviews, the guidance in scientific writing and the ongoing practical support for

our family.

Finally, my husband Shaun has been the driving force behind completion of this

thesis. This would not have been completed without him and it is to him - and our

children - that I dedicate this thesis.

10

STATEMENT OF CANDIDATE CONTRIBUTION

I, Helen Wilcox, hereby declare that:

This thesis contains published work and/or work prepared for publication, some of

which has been co-authored. The bibliographical details of the work and where it

appears in the thesis are outlined below.

Location in thesis: Components located in Methods, Results and Discussion

Further, I declare that:

This thesis contains no material which has been accepted for the award of any

other degree or diploma at any university or institution and;

to the best of my knowledge and belief, this thesis contains no material previously

published or written by another person, except where due reference is made in the

text of the thesis.

Student Signature

Helen Wilcox

23 December 2015

11

CHAPTER ONE: BACKGROUND

1. ACUTE CORONARY SYNDROMES IN AUSTRALIA

1.1 Acute coronary syndromes: prevalence and burden

Acute coronary syndromes (ACS) are a life threatening form of coronary heart disease

(CHD) and a leading cause of illness and death in Australia, with 69,900 Australians

experiencing an ACS in 20111. While mortality due to ACS is declining, due to better

control of CHD risk factors and the introduction of new treatments2-4, at least 10,000

Australians still die each year from ACS1. ACS also place a large burden on tertiary

hospital resources, with $8 billion of health care expenditure spent annually on

inpatient care of the condition5. Excluding the diagnosis in a patient with ACS

symptoms also commands resources; ACS symptoms are one of the most common

reasons for a patient to present to an emergency department (ED), even though 75 to

85% of these patients do not ultimately have a diagnosis of ACS6.

1.2 ACS: pathogenesis, classification and diagnosis

The spectrum of ACS includes unstable angina (UA), where atherosclerotic plaque

rupture leads to arterial occlusion and myocardial ischaemia, and acute myocardial

infarction (AMI), where ischaemia progresses to myocardial cell necrosis. Further

classification into ST elevation myocardial infarction (STEMI) and non-ST elevation

myocardial infarction (NSTEMI) is based on electrocardiographic (ECG) findings, as

shown in Figure 1.

The diagnosis of ACS is suggested by typical clinical features in the presence of CHD

risk factors, and confirmed by the presence of supportive ECG and biochemical

abnormalities7-9. The typical pain of ACS is a severe, retrosternal pressure or

heaviness which occurs at rest, is prolonged or recurrent, radiates to the back, neck or

arm, and is not relieved by sublingual nitrates8-10. Atypical presentations do occur, more

commonly in women, the elderly, and diabetics11,12 and are associated with delayed

presentations to medical care and missed diagnoses13,14. Such presentations include

pain that is unusual in location, being felt in the back, neck, arm, or epigastrium without

chest pain, or unusual in nature, being sharp or pleuritic. Atypical presentations also

include symptoms of syncope, presyncope, dyspnoea, nausea and vomiting. When

present as part of an atypical picture, these symptoms may overshadow the pain or be

present in the absence of pain8-10. Additionally, it is possible to have no symptoms at

all, with the diagnosis of ACS made on ECG or cardiac imaging11.

12

Figure 1: Classification of acute coronary syndromes

Presentation suggestive of acute coronary syndrome

ECG: ST depression

Non -ST elevation acute coronary syndrome (NSTEACS)

ECG: new ST elevation(or left bundle branch block)

ST elevation myocardial infarction

(STEMI)

Troponin positive:Non-ST elevation

myocardial infarction (NSTEMI)

Troponin negative:Unstable angina

(UA)

Troponin

Adapted from: White HD, Chew DP. Acute myocardial infarction. Lancet. 2008 Aug 16;372(9638)15.

Figure 2, Classification of acute coronary syndromes, p572. Multiple prospective studies have shown that clinical features alone are not reliable

predictors of ACS16-18, especially in older patients19. Additionally, for the majority of

patients with symptoms of ACS, an initial ECG will not show ST segment elevation20.

Consequently, biochemical investigations are required for diagnosis.

2. TROPONIN

2.1 The troponin complex

Cardiac troponin (cTn) is the main biomarker for the diagnosis of AMI9,11,21,22. It is a

structural protein complex in skeletal muscle and myocardial cells, comprising the

subunits TnT, TnI and TnC9. The three subunits are located along the length of actin

filaments in myofibrils and play an important regulatory role in the calcium-dependent

activation of muscle contraction, as shown in Figure 2.

In ACS, initial myocardial ischaemia results in release of the TnT and TnI subunits into

the blood. This release can be detected by biochemical assay as early as two hours

after the onset of ischaemia. The magnitude and duration of cTn elevation are

proportional to the severity of the myocardial injury23. Elevations in TnI levels can

persist for seven to ten days in patients with large AMI, probably due to the ongoing

release of cTn during myonecrosis and the subsequent remodelling process24.

13

Conversely, cTn elevation in smaller infarctions might last several hours only at a time,

due to its short circulating half-life and an absence of ongoing necrosis.

Figure 2: The troponin complex.

In a resting muscle fibre, the troponin complex inhibits actin–myosin interaction (a). Calcium release from intracellular stores causes a conformational change in the complex which allows tropomyosin bound along the actin strands to shift position and expose the myosin-binding sites (b). Myosin then binds to actin, actin and myosin slide past each other, and muscle contraction results. From: Reece JB Campbell NA. Campbell Biology: Australian Version 9th ed: Pearson Education Australia; 201125. Figure 50.29, The role of regulatory proteins and calcium in muscle fiber contraction, p1106.

2.2 Troponin in diagnosis and risk stratification of ACS

The diagnostic criteria for AMI, as defined by the Joint ESC/ACCF/AHA/WHF Task

Force for the Universal Definition of Myocardial Infarction11, comprise the presence of a

rise and/or fall in at least one cTn level above the 99th centile for the reference

population, combined with symptoms suggestive of ischaemia and findings supportive

of ischaemia on ECG, functional imaging and/or angiography (Table 1). A elevated cTn

level may be the only feature in addition to history that permits the diagnosis of

NSTEMI to be made, as the majority of patients experiencing a NSTEMI will have a

normal physical examination, and half will have a normal ECG8.

The use of a rise or fall in cTn levels, known as a dynamic change, allows detection of

new onset AMI with increasing values, and resolving AMI with decreasing values26. The

minimum change required, and whether an absolute or relative change is preferable,

varies between guidelines. A change of at least 50% is recommended by Australian

14

Table 1: Definition of myocardial infarction

The term acute myocardial infarction (MI) should be used when there is myocardial necrosis in a clinical setting consistent with acute myocardial ischaemia. Under these conditions any one of the following diagnoses meets the criteria for MI: 1. Detection of a rise and/or fall of cardiac biomarker values (preferably cardiac troponin [cTn] with at least one value above the 99th percentile upper reference limit [URL]) and with at least one of the following: • Symptoms of ischaemia • Development of pathologic Q waves in the electrocardiogram (ECG) • New or presumed new significant ST-segment-T wave (ST-T) changes or new left bundle

branch block (LBBB). • Identification of an intracoronary thrombus by angiography or autopsy • Imaging evidence of new loss of viable myocardium or new regional wall motion

abnormality. 2. Cardiac death with symptoms suggestive of myocardial ischaemia and presumed new ischaemia ECG changes or new LBBB: • Death occurred before cardiac biomarkers were obtained, or before cardiac biomarker

values would be increased. 3. Percutaneous coronary intervention (PCI)-related MI: • Elevation of biomarker values (cTn is preferred) >5 x 99th percentile URL) in patients with

normal baseline values (<99th percentile URL) or a rise of values >20 percent if the baseline values are elevated but stable or falling.

• In addition, either (i) symptoms suggestive of myocardial ischaemia, (ii) new ischaemic ECG changes or new LBBB, (iii) angiographic loss of patency of a major coronary artery or a side branch or persistent slow- or no-flow or embolization, or (iv) imaging demonstration of new loss of viable myocardium or new regional wall motion abnormality are required.

4. Stent thrombosis associated with MI: • Detected by coronary angiography or autopsy in the setting of myocardial ischaemia and

with a rise and/or fall of cardiac biomarkers, with at least one value above the 99th percentile 5. Coronary artery bypass graft surgery (CABG)-associated MI: • Elevation of cardiac biomarker values >10 x 99th percentile URL in patients with normal

baseline cTn values. • In addition, either (i) new pathologic Q waves or new LBBB, (ii) angiographic documented

new graft or native coronary artery occlusion, or (iii) imaging evidence of new loss of viable myocardium or new regional wall motion abnormality.

Adapted from: Thygesen K, Alpert JS, Jaffe AS, Simoons ML, Chaitman BR, White HD, et al. Third Universal Definition of Myocardial Infarction. Circulation 2012 Oct; 126(16): 2020-2035. Definition of myocardial infarction, p2022. authors27 (based on reported intra-individual variation of up to 46%) and a change of at

least 20% is required elsewhere11, though a greater relative increase than this is

suggested if the initial value is close to the upper limit of normal28. The exact algorithm

used can vary between laboratories depending on the precision of their assay29.

In addition to its role in diagnosis, cTn levels contribute to the assessment of risk of

adverse outcomes in patients with ACS, with increasing cTn levels correlating with a

proportional increase in mortality9,21,30-32. Such adverse outcomes are not infrequent,

with a 4.5% rate of in-hospital mortality for patients with AMI and a 5.1% rate of

recurrent AMI. Figures are even higher in patients with STEMI33. Overall, the rate of an

in-hospital major adverse cardiac event as a consequence of ACS (such as death,

cardiac arrest, recurrent MI, worsening heart failure, major bleeding or stroke)

15

approaches 30% for STEMI and 20% for NSTEMI33. Patients with UA also have an

increased risk of cardiac death and subsequent MI despite the lack of myonecrosis9.

2.3 Troponin in non-ACS contexts

cTn can be elevated in clinical contexts apart from ACS, and elevation in these

situations also confers a higher risk of adverse outcomes. Firstly, patients with stable

CHD and elevated cTn are at higher risk of major adverse events34-37. Additionally,

patients with normal coronary arteries on angiography and elevated cTn retain a higher

risk of mortality than similar patients with a normal cTn level26. The mechanism of

increased risk in these settings may be coronary artery vasospasm, embolisation or

dissection, or early aggressive antithrombotic therapies removing visible thrombus by

the time the vessels are demonstrated angiographically15,38.

Non-coronary conditions which increase myocardial oxygen requirement or decrease

oxygen supply can also cause cTn elevation in the absence of atherosclerosis, as

listed in Table 2. In some of these disease states the cTn elevation may be stable and

chronic, and use of dynamic cTn levels will help with their differentiation from MI.

However, the more acute conditions such as sepsis may also produce a dynamic cTn

level due to acute cardiac injury in the absence of MI39. The presence of heterophile

antibodies in certain individuals and assay calibration errors may also lead to an

elevated result11.

For these reasons, the diagnosis of ACS relies on a supportive clinical picture in

addition to cTn elevation, and in the absence of a history suggestive of ACS or of CHD

risk factors other causes of the cTn rise should be sought.

Table 2: Non-coronary causes of elevated cTn

Ischaemic Non-ischaemic Non-ACS causes of MI Cardiac Systemic Hypoxia Heart failure Pulmonary embolism Global ischaemia Infection (e.g. viral

cardiomyopathy) Anthracycline toxicity

Hypoperfusion Inflammation (e.g. myocarditis) Trauma (e.g. chest wall injury) Cardiothoracic surgery Trauma (e.g. surgery) Renal failure Ablation procedures Sepsis Malignancy Stroke Stress cardiomyopathy Subarachnoid haemorrhage Infiltrative diseases

Adapted from: Newby LK, Jesse RL, Babb JD, Christenson RH, De Fer TM, Diamond GA, et al. ACCF 2012 Expert Consensus Document on Practical Clinical Considerations in the Interpretation of Troponin Elevations - A Report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents. J Am Coll Cardiol 2012 Dec 11; 60(23): 2427-2463. Figure 1: Conceptual Model for Clinical Distribution of Elevated Troponin

16

2.4 Delays in cTn testing

Early, rapid diagnosis of ACS is vital40. The overall mortality rate for ACS within one

month of the event is approximately 50%, with half of these deaths occurring within two

hours of the event10. Despite this need for early diagnosis of ACS, there are multiple

delays inherent in ACS diagnosis using cTn testing. Standard laboratory based cTn

testing has been limited by a lack of sensitivity for diagnosis of ACS in the early hours

following infarction, with elevation not detectable for four to six hours8,45. Consequently,

a cTn result is only considered to be sufficiently sensitive when tested at least eight

hours after onset of symptoms27,46.

There are also delays associated with sample collection and transportation to the

laboratory. Such delays are minimal for patients who present to hospitals with co-

located laboratories, but can be significant when the laboratory is at a distant site.

Once the sample has been transported to the laboratory, sample preparation, analysis

and result reporting need to occur before a result is available47. For patients with

possible ACS, the time taken for this sequence of events should be no more than 60

minutes, according to Australian and international guidelines27,48.

Point of care testing (POCT) devices using cTn testing have much shorter turn-around

times than laboratory based testing regimes. POCT results are available within 15 to 26

minutes, due to elimination of transportation time and minimal sample preparation time,

as most devices use whole blood26. A POCT strategy might alleviate some issues with

delay and has been shown in the ED setting to reduce ED length of stay for patients

with a normal cTn result, when used in conjunction with clinical risk scoring systems6,49-

51. The downside of POCT is the limited sensitivity of the assays, with some assays

having lower accuracy than laboratory based tests52,53. Additionally, analytical

variability exists between POCT cTn assays, posing a risk of misinterpretation of

results if the performance of the particular assay used is not taken into consideration.

17

2.5 Monitoring during cTn testing

Until a cTn result is known to be negative, major Australian and international

guidelines recommend continuous cardiac monitoring for patients with suspected

ACS7,27,54. The rationale for monitoring includes the need to identify evolving ECG

changes that would indicate eligibility for reperfusion. In addition, abnormalities of the

ST segment on ECG also provide prognostic information, being independently

associated with an increased risk of adverse outcomes55. Cardiac monitoring may also

identify complications of evolving infarction such as cardiac arrhythmias and

hemodynamic instability.

2.6 Highly sensitive troponin in ACS diagnosis

Recently, highly sensitive troponin (hsTn) assays have been developed that that have

10 to 100 fold lower limit of detection of ACS compared with standard methods of

testing56,57. This improved sensitivity lowers morbidity and mortality rates for ACS

patients with detectable hsTn levels, presumably due to improved access to

reperfusion therapy with earlier diagnosis58. When incorporated into established risk

stratification scoring systems, hsTn can reliably rule out MI as early as six hours after

symptom onset27, resulting in shorter emergency department transition times for those

with normal hsTn levels who can safely be designated as being at low risk of adverse

outcomes57,59-61. This approach is also cost effective when compared with testing in the

standard eight hour time period62. hsTn testing early after presentation is not yet

included in National Heart Foundation guidelines27 but its inclusion is predicted in the

future63,64.

The lower limit of detection of hsTn assays comes at the cost of reduced specificity,

with measurable concentrations of hsTn detectable in significant numbers of

asymptomatic individuals28,65. Conventional assays have been shown to detect

circulating hsTn in 0.7% of the community population without known CHD66, with newer

hsTn assays increasing this to 50-66% of the population aged over 6537,67 and up to

80% of a younger age group68. In addition, other factors affect the interpretation of

hsTn results, such as individual biological variation in cTn levels over time26, the

heterogeneity of assays available26,69,70 and whether an appropriate reference range is

that of a younger healthy population or that of an aged matched population71.

As a result, the use of hsTn continues to be a topic of debate28,72-77.

18

3. TROPONIN AND ACS IN AUSTRALIAN GENERAL PRACTICE

3.1 Clinical diagnosis of ACS in general practice

The challenge of ACS diagnosis exists in primary care as well as in the tertiary setting.

In Australia, primary care presentations with symptoms associated with ACS such as

chest pain are common, with 15% of patients experiencing an acute MI first contacting

their general practitioner with their symptoms78. The number of such consultations is

substantial: there are over 126 million general practice all cause encounters each year

in Australia and at least 1% of these presentations involve chest pain as a reason for

the encounter79. The Bettering the Evaluation And Care of Health (BEACH) Program80

reports that:

i. 8% of patients seeking GP care have established CHD with a high risk of ACS;

ii. 3.6% have a history of acute myocardial infarction in the last three years, and;

iii. 60% of patients in general practice have at least one CHD risk factor.

The typical ACS presentation of acute prolonged or recurrent central chest pain with

unequivocal ischaemic changes on ECG in a patient with multiple CHD risk factors is

well known to GPs. The decision to refer to hospital for a definitive diagnosis, rather

than proceed with outpatient investigations, is usually clear in these patients81. In less

typical presentations, ACS diagnosis in primary care is not always straightforward, as

there are limitations to the clinical tools that can be used within the consultation to

assess the risk of ACS. Evidence from systematic reviews shows that in the pre-

hospital population, signs and symptoms alone are neither sensitive nor specific for the

diagnosis of ACS81-84. Additionally, risk stratification tools such as the Thrombolysis in

Myocardial Infarction (TIMI)85 and the Global Registry of Acute Coronary Events

(GRACE)86 scores were developed using secondary care populations which limit their

generalisability to patients in primary care.

GPs seem to use a combination of strategies in diagnosis of serious illness and in

managing diagnostic uncertainty: their initial impression or “gut feeling”, the

combination of clinical factors which are of limited value in isolation but highly sensitive

when applied together, and development of a safety netting strategy87,88. Attempts to

formalise this approach have led to the development of clinical prediction rules for ACS

in the general practice setting. These report negative predictive values of 94.8% to

97.9%89-91 for the diagnosis of ACS. However, guidance directed at Australian primary

care doctors suggest their validation has been limited and their accuracy cannot be

relied upon63. The inclusion of ECG findings does not always improve accuracy of

19

diagnosis. GPs do not always interpret key findings on ECGs correctly, with estimates

of correct interpretation ranging from 59 to 70%92,93.

3.2 Benefits and limitations of cTn testing in general practice

Given the limitations of the above clinical tools in primary care, cTn testing in primary

care should theoretically improve accuracy of diagnosis of ACS. Even so, cTn

interpretation in primary care can be as complex as in the hospital setting. Appropriate

use requires an understanding of its sensitivity and specificity as well as consideration

of technical and practical issues.

According to the American College of Cardiology, “the concept of appropriateness, as

applied to health care, balances risk and benefit of a treatment, test, or procedure in

the context of available resources for an individual patient with specific

characteristics”94. Figure 3 describes the potential effects of differing cTn test results on

GP management decisions.

Figure 3: Potential effects of GP-initiated cTn test results on management

GP provisional Dx plus referral plan

Patient presents with ACS Sx

Test

+ ve

HospitalNot appropriate: no effect, risk of

delay, AO

Unlikely scenario

Appropriate: beneficial (if not false –ve)

Appropriate: beneficial (if not false +ve)

Appropriate: beneficial (if not false +ve)

Unlikely scenario

Not appropriate: no effect,

resource issue

PatientPresentation Test Result Referral Effect

of test

No hospital

Hospital

No hospital

Hospital

No hospital

Hospital

No hospital

- ve

+ ve

- ve

Provisional Plan

Likely ACSHospital

TestUnlikely ACSNo hospital

Box

1

2

3

4

5

6

7

8

Not appropriate: no effect, risk of

delay, AO

+ve = positive, –ve = negative, AO = adverse outcome, Dx = diagnosis, hospital = emergency department evaluation.

20

Only some of the decisions in Figure 3 can be considered appropriate. When this

definition is applied to cTn, appropriate use of cTn should:

(a) avoid the risk of adverse outcomes of ACS which may occur if cTn testing delays

hospital referral (Boxes 1 and 3)

(b) benefit the patient by excluding ACS in those at low risk of ACS (Box 4)

(c) benefit the patient by allowing diagnosis of ACS and expedited referral to hospital

where ACS was initially not suspected due to atypical presenting features (Box 5)

(d) benefit the patient by providing valuable diagnostic and prognostic information in

patients not suitable for admission, such as those in residential care or with significant

comorbidities (Box 6)

(e) be interpreted with an understanding of specificity and sensitivity (Boxes 4, 5 and 6)

(f) conserve emergency department resources by avoiding attendance at hospital

emergency departments 95 (Box 8). It is worth noting that use in this context may not

translate to a substantial conservation of ED resources, as the total burden of GP

patients in ED is reportedly overestimated96.

Multiple authors have expressed concern that cTn is used inappropriately by GPs.

These concerns relate to delay in hospital referral, an over-interpretation of positive

results and an overreliance on negative results21,94-99.

3.2.1 Delay in hospital referral

The first situation where cTn use may be inappropriate is when hospital referral is

required regardless of test outcome. In this context, cTn testing may cause

unnecessary delay (Figure 3, Boxes 1 and 3). Patients with ACS whose first medical

contact is a community physician rather than emergency services experience greater

pre-hospital delay14,97-99 especially if presenting with atypical symptoms100,101. There is

a risk of misdiagnosis102, greater haemodynamic instability103 and delayed access to

thrombolysis104,105 as well as increased mortality in these patients106.

Adverse cardiac events of the ACS may occur during this period of delay, and urban

GPs do not always have access to the equipment and other resources needed to

detect and manage complications such as cardiac arrhythmias. Access to continuous

ECG monitoring is limited to the tertiary inpatient setting8,9 and is not routinely required

in urban primary care. Accredited GP practices in Australia do not need an ECG

machine for accreditation purposes, nor are they required to have an onsite defibrillator

for management of cardiac arrhythmias107.

21

Theoretically, point of care troponin testing (POCT) in primary care, as used in the ED

setting, may go some way towards resolving the issue of delay in ACS diagnosis. Such

a strategy would require validation on an Australian primary care population, as the

prevalence of ACS is lower than in the ED setting, thereby affecting the predictive

value of the test108. There are also problems with POCT assay standardization, staff

training and a lack of Medicare subsidies, which limit use in primary care52,109-111.

It is worth noting that if patients present to GPs late in the course of the ACS (after the

48 hour period of maximum risk for short term adverse outcomes) there may be no

increase in adverse outcomes. In late-presenting patients in whom invasive

management is less time-critical, delay due to cTn testing may not be relevant.

3.2.2 Over-interpretation of positive results

The type of chest pain seen in primary care differs from that seen in ED. The

prevalence of ACS in primary care populations, and consequently the pretest

probability of ACS, is lower than in secondary care. Pain due to a musculoskeletal

cause is the most common diagnosis for primary care chest pain, with only 10-34% of

presentations having CHD as their cause112-114.

Additionally, non-coronary conditions that cause detectable cTn levels are prevalent in

general practice and falsely positive test results are therefore likely115. Increasing use

of hsTn will add to uncertainty, with many primary care patients returning elevated

hsTn results having non-ACS conditions that do not warrant ED attendance. Detectable

hsTn values may be solely a result of recent exercise or increasing age, with positive

results in the very old being the norm71. Without laboratory provision of age-related cut-

off values, such results may trigger unnecessary intervention.

3.2.3 Overreliance on negative results

Recent Australian literature has suggested that GPs have incomplete understanding of

the need for serial cTn testing, as serial testing is underutilised95. Even with serial

testing, negative results in the context of chest pain do not obviate the need for

second-line evaluation. ED patients with chest pain, normal ECGs and normal cTn

levels remain at risk of adverse outcomes, with an outcome of ACS, urgent coronary

revascularisation or death of cardiac origin in 1 of 40 such cases at 30 days after an

episode of chest pain, and in 1 of 14 cases at 6 months116.

A further complication for GPs is the variability of cTn assays. Available tests have

different sensitivities and use differing reference populations to define the cutoff value

at the 99th percentile52,117. This is particularly relevant when comparing serial results

22

from different laboratories, as a significant rise or fall in cTn values may not be

appreciated if GPs use multiple laboratories for one patient episode.

3.3 Practical issues in cTn testing in primary care

There are practical implications of cTn testing in primary care. While a request for

cardiac biomarkers is regarded as urgent by the laboratory, it may be some hours

before a result is available, potentially only after the GP’s surgery is closed. Difficulties

then arise if the requesting doctor is not contactable, as it falls to the pathologist, a

locum service or a colleague of the requesting GP to contact the patient and arrange

further investigation. A 2005 Coronial Inquest examined a patient death following a cTn

test ordered in general practice that was not attended to in a timely manner due to the

result being notified after office hours118. The conclusion was that there was a failure of

systems in place at the medical centre for the patient’s test results to be accessed,

assessed and appropriate action taken118. While systems can be established within

group practices to allow successful follow up of abnormal life threatening results

outside opening hours, and accreditation of general practices requires after hours

cover to be available107, not all practices may consistently operate to this standard.

Pathology providers across Australia recognize the potential risks to patients in

undergoing cTn testing in primary care. Historically, some Queensland laboratory

collection centres will not collect blood samples for outpatient cTn testing, referring

patients to ED instead (Dr Narelle Hadlow, pers.comm.). In Western Australia major

pathology providers have stressed the need to exercise caution in ordering cTn on

community patients, reminding GPs of the delays inherent in non-hospital cTn testing

and stating that in cases with a high index of suspicion it may be more appropriate to

refer patients to a tertiary care facility for assessment119-121.

3.4 Estimates of frequency of cTn testing in primary care

Despite these limitations, cTn testing for the diagnosis of ACS remains available for

use by GPs in the primary care setting under the Australian Medicare Benefits

Schedule122. Across Australia in 2012, 481,322 cTn tests were ordered outside public

hospitals123 at a cost to Medicare of over $10 million. While this figure includes tests

performed in private hospital and by specialists other than GPs, the majority are likely

to have been requested by GPs, given that pathology tests requested by GPs account

for 70% of Medicare pathology services124.

23

4. GUIDELINES FOR PRIMARY CARE USE OF TROPONIN

4.1 Australian guidelines

Recent reviews have made suggestions on the use of cTn testing in general

practice63,95. Authors suggest that a single cTn test may be used in general practice to

exclude the possibility of AMI in asymptomatic patients whose symptoms resolved at

least 12 hours prior to the test, so long as they have no features placing them at high

risk of ACS, and a normal ECG. Serial tests are advised in patients presenting within

12 hours of symptom onset who are at low risk of ACS and/or have atypical symptoms.

Despite this, there is no formal guideline addressing appropriate use of cTn testing

directed at Australian GPs from the Heart Foundation, the Royal Australian College of

General Practitioners (RACGP) or the Royal College of Pathologists Australasia

(RCPA). The RACGP-supported point of care database – Dynamed125– does not

address how to use cTn in primary care as opposed to the hospital setting. A recent

review in the RACGP’s own journal126 advises cTn testing as part of the investigation of

possible ACS but makes no mention of the attendant risks and limitations. The RCPA

Manual speaks in general terms of the complementary nature of laboratory

investigations for diagnosis of MI but does not discuss their application in primary

care127. Locally, the Western Australian Health Department ACS Model of Care

acknowledges that some patients with possible ACS may present to GPs but it does

not advise GPs on management other than hospital referral40.

4.2 International guidelines

In the international literature, few of the major European or United States guidelines on

ACS contain recommendations to GPs about cTn testing. Authors of such guidelines

include the Scottish Intercollegiate Guidelines Network128, the National Institute for

Health and Care Excellence129, the European Society of Cardiology7, the National

Academy of Clinical Biochemistry and the American College of Cardiology

Foundation130. The Guidance on chest pain of recent onset from the National Institute

for Health and Care Excellence129 advises obtaining a blood cTn level only if a recent

ACS is suspected in people whose last episode of chest pain was more than 72 hours

ago and who have no complications of ACS such as pulmonary oedema, and that

clinical judgement should be used to decide whether referral is necessary and how

urgent this should be.

Only one other guideline is directed at GPs to inform safe practice in cTn testing,

produced by NHS Quality Improvement Scotland131. It recommends that cTn should not

24

be measured in primary care, except where the clinical episode of pain is between 24

and 72 hours earlier, or where the patient has other medical conditions precluding

admission to hospital. This advice is at odds with that of the Australian authors63,95.

Other international guidelines for GPs to use in diagnosis of chest pain intentionally

exclude cTn, citing its perceived limited value in primary care89.

5. CONCLUSION

In summary, ACS is common and carries a high risk of death and adverse outcomes. While cTn testing is central to the diagnosis of ACS, there are a number of pitfalls and

practical considerations in its interpretation. POCT and the use of high sensitivity cTn

do not entirely resolve these issues. Urban GPs frequently encounter patients with

symptoms suggestive of ACS, and cTn testing is frequently performed in the

community to investigate these presentations. However, urban general practice is not

structured to minimise delays in cTn testing and does not have the capabilities to

detect and manage complications of ACS. It has also been suggested that GPs have

incomplete knowledge of the limitations of cTn testing.

It is not known if the above issues with GP-initiated cTn testing actually result in

adverse outcomes of ACS. If the test is ordered often in an unsupported setting on

patients at high risk of ACS, adverse outcomes may eventually occur. Alternatively,

cTn testing in primary care might be beneficial if it permits diagnosis of ACS that would

have otherwise been overlooked on clinical grounds. Additionally, if GP-initiated cTn

testing rules out ACS, use of the test might alter GP management, reduce the number

of referrals to ED and so conserve ED resources.

This study aims to explore these issues by examining the use of cTn testing for ACS

diagnosis in primary care. The study will describe a population of patients who undergo

cTn testing in general practice and assess whether they have a lower risk of ACS than

those who present directly to ED. GP knowledge of cTn testing’s use and limitations

and the influence of cTn testing on GPs’ diagnosis and management will be explored.

Finally, the outcomes of patients who undergo cTn testing in primary care will be

examined.

25

AIMS

1. To describe the clinical characteristics of a patient group who undergo cTn

testing in primary care, specifically:

i. The nature and duration of presenting symptoms;

ii. The coronary risk status of the patient prior to testing;

iii. An assessment of whether patients in the group are clinically different and at a

lower risk of ACS to those who present directly to emergency departments with

ACS symptoms.

2. To examine GPs’ knowledge of cTn’s use and limitations of the test in general

practice, specifically the risk of false positive and negative results.

3. To assess the effect of the test result on the GP’s:

i. estimation of likelihood of ACS prior to test ordering and following receipt of

results;

ii. management of that patient prior to test ordering and following receipt of

results.

4. To document the outcomes of patients in this group who underwent cTn testing

in general practice. Specific outcomes sought would be:

i. Hospital presentation or admission with cardiac symptoms or diagnoses;

ii. Hospital procedures relating to ACS;

iii. Occurrence of significant events representing adverse outcomes of ACS;

iv. Delay in diagnosis of ACS.

HYPOTHESES

1. Patients undergoing cTn testing in primary care are clinically different to those who

present directly to emergency departments

2. CTn testing in primary care does influence a GP's management of their patient.

3. CTn testing in primary care is associated with adverse patient outcomes.

26

CHAPTER TWO: TROPONIN TESTING FOR ACUTE CORONARY SYNDROME IN PRIMARY CARE - A SYSTEMATIC REVIEW

1. RATIONALE

GPs have a key role in assessment of patients with symptoms suggestive of ACS.

They must integrate clinical findings and investigation results, which may include cTn,

in order to determine the need for hospital referral or to exclude ACS7-9. Because cTn

interpretation is complex, GPs require an understanding of its sensitivity and specificity

as well as of its technical and practical considerations in order to use cTn testing

appropriately. It is important to know if GPs have this understanding and are able to

use cTn appropriately, as inappropriate use could place patients at risk of adverse

outcomes.

2. OBJECTIVES 1. To describe the outcomes of patients who presented with ACS symptoms in

general practice and had cTn testing performed;

2. To examine GPs’ understanding of cTn testing’s limitations, and the influence of

cTn test result on GP management.

3. METHODS

3.1 Theoretical and methodological approach

The standard methodology for systematic reviews was followed according to guidelines

published by the National Health and Medical Research Council132 and the York Centre

for Reviews and Dissemination133. The development of a theoretical model to assist

with the search strategy and the process for narrative synthesis were guided by the

ESRC Methods Program Guidance on the Conduct of Narrative Synthesis for

Systematic Reviews134. Findings are reported according to PRISMA (Preferred

Reporting Items for Systematic Reviews and Meta-Analyses) guidelines135,136. Methods

of the analysis were pre-specified and documented in a protocol.

3.2 Eligibility criteria

Table 3 presents the inclusion criteria for the review with reference to standardised

data definitions used in Australasian ACS research137.

27

Table 3: Eligibility criteria for systematic review

Criteria Included Excluded Participant related: Definition of disease/condition

1.ACS and synonyms (Acute myocardial infarction, STEMI, NSTEMI, Unstable angina)

Non-ischaemic heart failure.

2.Ischaemic heart disease Presentation (Note 1)

Symptom(s) of ACS without a previous ACS diagnosis for this episode of symptoms

No symptoms; cTn used as screening test

Demographic factors All None

Setting Urban primary care 1.Wholly presenting to Emergency department

2.Hospital inpatient at time of testing

3.Populations where time taken to perform test is greater than time taken to present to inpatient setting

4.Test ordered by paramedical services

Intervention related:

Intervention 1. cTn (including TnI, TnT, hsTn) requested by GP ECG only

2. Laboratory or point of care testing Control No cTn test ordered by GP

Outcome related: Survival 1. Death from cardiovascular cause (Note 2) 2. Death of uncertain cause

3. Cardiac arrest Adverse events 1. Cardiogenic shock (Note 3) 2. Ventricular arrhythmia within 48 hours

3. High degree AV block within 48 hours 4. AMI within 30 days diagnosed on ECG 5. STEMI within 30 days diagnosed on ECG 6. NSTEMI within 30 days diagnosed on ECG

7.AMI plus NO Emergency revascularization procedure NOR Urgent revascularization procedure NOR thrombolysis

Delay (Note 4)

Delay greater than 30 minutes or 90 minutes depending on time from medical contact to PCI

Beneficial effects Late or atypical presentation PLUS any of: (Note 5) 1.ACS or synonyms

2.Emergency revascularization procedure 3.Urgent revascularization procedure 4.Elective revascularization procedure

Note 1: Symptoms consistent with possible ACS include acute chest, epigastric, neck, jaw or arm pain or discomfort or pressure without an apparent non-cardiac source. Atypical presentations of ACS may include back pain, sharp and pleuritic pain and the following symptoms in the absence of pain: dyspnoea, palpitations, nausea, vomiting, diaphoresis, fatigue, syncope or presyncope137. Note 2: The relevant time period in survival is the first 48 hours where inpatient monitoring is indicated. Note 3: Australian consensus guidelines list of important outcomes to monitor in patients with ACS137. The relevant time period is the first 30 days after ACS where there is the greatest risk of progression to MI or the development of recurrent MI or death. Following this phase most patients with ACS resume a clinical course similar to that of patients with stable coronary disease9,138. Note 4: In general, the maximum acceptable delay from first medical contact to balloon inflation (marking the commencement of PCI) is 60 minutes if a patient presents within 1 hour of symptom onset; or 90 minutes if a patient presents within 1-12 hours of symptom onset. PCI is still considered in those who present more than 12 hours after onset of symptoms if there are ongoing symptoms, haemodynamic instability or a contraindication to fibrinolysis.8 Note 5: Late presentation is defined as greater than 48 hours after symptom onset as this is window where inpatient monitoring is indicated.

28

3.3 Information sources

An initial search was conducted for existing or ongoing reviews of cTn testing in

primary care in the following registers of systematic reviews:

1. Database of Abstracts of Reviews of Effects (DARE)139

2. Cochrane Database of Systematic Reviews (CDSR)140

3. National Institute for Health and Clinical Excellence (NICE)141

4. NIHR Health Technology Assessment (NIHR HTA)142

5. National Guidelines Clearinghouse (NGC)143

6. Scottish Intercollegiate Guidelines Network (SIGN)144

No existing systematic reviews were identified in this process.

3.4 Search

A search strategy was then designed in consultation with a medical librarian. The

search involved the following database list:

1. Web of Science

2. MEDLINE

3. Evidence Based Medicine Reviews Multifile Database:

3.1. ACP Journal Club

3.2. Database of Abstracts of Reviews of Effects

3.3. Cochrane Central Register of Controlled Trials

3.4. Health Technology Assessment

3.5. Cochrane Database of Systematic Reviews

3.6. National Health Service Economic Evaluation

3.7. Cochrane Methodology Register

4. Embase

5. Scopus

6. TRIP

Table 4 shows the final list of potential search terms.

29

Table 4: List of terms used in electronic search strategy

ACS symptoms Pre-hospital care Troponin General Practice Delay

Cardiovascular diseases

Emergency Medical Services

Point of care testing

General Practitioner

Diagnostic Errors

Acute Coronary Syndrome

Pre-hospital Care Troponin I Primary Health Care

Misdiagnosis

Coronary Artery Disease

Resuscitation Troponin T Family Physician

Time Factors

Coronary Disease Troponin Family Practice

Delay

Coronary Thrombosis General Practice

Unstable Angina Primary Care Myocardial Infarction Angina Pectoris Chest Pain Arrhythmia Fatigue Syncope

The search strategy for relevant papers was customised for each database using a

combination of MeSH and free text non-indexed keywords including truncations and

wildcards. The search was piloted with the aim of obtaining four papers known to be

relevant for the review. The strategy was refined following this initial exercise.

Related links, citing articles and reference lists for key papers were reviewed. Hand-

searching was performed in full text journals in the fields of primary care, cardiology,

emergency medicine and pathology. Reports from the Australian Institute of Health and

Welfare and the Heart Foundation were assessed for relevance. Grey literature

sources listed in the Cochrane Handbook were reviewed. Reference tracking and

citation tracking using Scopus, Medline, Web of Science and Google Scholar were

performed to complete the search.

The search was limited from January 1990 to December 2013, 1990 being the first year

that the use of the cTn assay in diagnosis of ACS was described. No language or

document format restrictions were applied. The full search strategy for each database,

lists of journals hand-searched and grey literature sources are included as an

Appendix.

30

3.5 Study selection

A flow diagram of the study selection process is shown in Figure 4 (see Section 4.1

Study selection). Search results were merged and duplicates removed using EndNote

X6 reference management software145. Each title and abstract was examined by one

researcher (HW) to exclude irrelevant reports. Full text versions of remaining reports

were examined in full using the eligibility criteria in Table 3. Reports of unclear

significance were discussed with the other researcher (JE) and consensus obtained on

inclusion in the review.

3.6 Data collection process

A data extraction form was designed for the specific purpose of the review. The list of

items included was derived from the Cochrane Handbook of Systematic Reviews of

Interventions146. It was piloted and refined further using two papers that were highly

likely to be included and two of uncertain relevance.

3.7 Data items

Data were extracted from each included paper on: 1. Source (lead author, title, year published, citation)

2. Aims

3. Methods (design, setting, duration, year performed, study inclusion and exclusion

criteria)

4. Participant characteristics - patients (number, age, sex, symptoms, timing of

presentation, diagnosis)

5. Participant characteristics – GPs (number)

6. Intervention (Troponin T or I, laboratory based or POCT, number of tests

performed, timing of tests relative to symptom onset)

7. Quality assessment

8. Results of cTn test

9. Patient outcomes (hospital referral, hospital admission, survival, adverse events,

delay, beneficial effects)

10. Management before and after test result

3.8 Risk of bias in individual studies

All included studies were assessed for quality according to the Critical Appraisal Skills

Programme UK (CASP) checklist for cohort or qualitative research as applicable to

design of that study147. The NHMRC criteria for quality appraisal132 were also

considered in the formulation of results. Results of this assessment are summarised in

Table 6.

31

Due to the small volume of relevant studies retrieved, no studies were excluded on

methodological grounds. A grading system was used to rank the methodological quality

and relevance of each study according to the structured appraisal approach described

by Dixon-Woods148 and endorsed by the Cochrane Handbook146.

A search was undertaken for any published protocols for each study. None were found.

3.9 Synthesis of results

Given the diverse evidence types in the included papers a narrative synthesis

approach was used148 employing the framework described in the UK ERSC Guidance

on the Conduct of Narrative Synthesis in Systematic Reviews134. Steps in the

framework include:

1. Developing a theoretical framework

2. Developing a preliminary synthesis of findings of included studies

3. Exploring relationships within and between studies

4. Assessing the robustness of the synthesis.

32

4. RESULTS

4.1 Study selection

Figure 4 describes the study selection process. A total of seven studies were identified

for inclusion in the review. There was marked heterogeneity of these studies which

limited the ability to undertake meta-analysis, prompting a narrative synthesis approach

to analysis.

Figure 4: Flow diagram of study selection process.

4957 records identified through database searching

244 additional records identified through other sources

4707 records after duplicates removed

4207 records excluded

493 full-text articles excluded, with reasons:

No troponin test: 153GP not performing test: 202Non-ACS presentation: 35

No original data: 103

4707 records screened

500 full-text articles assessed for eligibility

7 studies included in systematic review:Objective 1 and 2: 3 studiesObjective 1 only: 2 studiesObjective 2 only: 2 studies

A description of included studies is provided in Table 5. Table 6 presents an

assessment of the risk of bias in each study.

33

Table 5: List of included studies

Study Aim Method Participants Intervention Follow up Country Setting Source of outcome

information Exclusion criteria Number Age Sex Symptom Number

of GPs Country

Aldous (2012)

To assess utilisation of troponin testing in the community

New Zealand (NZ)

Laboratory audit

National health events database

Admission to hospitals outside NZ

2662 tests 2575 patients

63 (51-74)

M 45%

NR NR Laboratory based Troponin I Standard sensitivity

6 months

Mann (2006)

To document clinical circumstances and medical outcomes where troponin is used to assess chest pain in general practice

New Zealand

Laboratory audit

GP survey Test added by laboratory and not requested by GP

433 tests 278 tests with survey data Number of patients NR

NR NR Chest pain in 245/278 surveys NR in 23/278 surveys

201 Laboratory based Troponin T Standard sensitivity

3 days

Planer (2005)

To evaluate the diagnostic value of troponin T kit testing in the community setting for the assessment of patients presenting with chest pain

Israel GP clinics GP diagnosis prior to test result, emergency physician diagnosis for positive tests, patient phone call follow up 2 months after study entry for all test results.

Age <30 years <20 consecutive mins chest pain Interval symptom onset to presentation <8hours or >6days Renal failure, ECG ST elevation ACS or revascularisation within previous 2 weeks


59 (+/- 14)

M 58%

Chest pain in 104/349 patients

NR POCT Troponin T

2 months

Tanner (2006)

To study behaviour of patients prior to admission to hospital with symptoms of ACS

New Zealand

Coronary care unit inpatients

Patient recall of time intervals

Patients transferred from other centres Language barrier Patients discharged or died prior to interview


65 (32-88)

M 69%

Chest pain in 91/100 patients

NA Laboratory based Troponin T Standard sensitivity

NR – data obtained during hospital admission

Tomonaga (2011)

To analyse the diagnostic accuracy of POCT in general practice for diagnosis of ACS, heart failure and thromboembolic events

Switzerland GP clinics GP requesting test Interval symptom onset to presentation >5days Anticoagulant therapy Renal failure Cancer therapy Practices within 8-10km of Zurich where rapid laboratory troponin results available

147 tests 151 controls

65 (+/- 16)

M 58%

Chest pain in195/369 patients

79

POCT Troponin T

3 weeks

34

Table 5(cont’d.): List of included studies

Study Aim Method Participants Intervention Follow up Country Setting Source of outcome

information Exclusion criteria Number Age Sex Symptoms Number of GPs

Law (2006)

To investigate GP knowledge and use of troponin testing in primary healthcare

Wellington region, New Zealand

GP clinics GP survey NR NA NA NA NA 299 30 GPs rural

NA NA

Sodi (2007)

To undertake an audit of troponin tests requested by GPs with a view to developing an informed strategy regarding assay provision and reporting of results

Liverpool, United Kingdom

Laboratory audit

GP telephone interview

NR 18 tests 16 tests with interview data

NR NR Chest pain 9/16

16 Laboratory based Troponin T Standard sensitivity

1 week

Tandjung (2012)

To survey infrastructure and the performance in diagnostics and therapy of cardiologists and GPs who encounter ACS in an outpatient setting

Switzerland GP and cardio-logist clinics

Postal survey NR NA NA NA NA 467 GPs 35 cardio-logists

NA NA

GP = general practitioner; NR = not recorded; POCT = point of care troponin; ECG = electrocardiogram; ACS = acute coronary syndrome; NA = not applicable

35

Table 6: Assessment of risk of bias in included studies using CASP tools

Cohort Study Appraisal Study Clearly

focused issue

Appropriate method

Recruitment acceptable

Exposure accurately measured

Outcome accurately measured

Confounders identified, accounted

for

Follow up long,

complete

Result presentation

Result precision

Results believable

Result applicability

Results fit with other evidence

Quality assessment

Aldous Y Y Y Y U N Y U U U Y Y SAT

Mann Y Y Y Y N U U U U U Y Y ?

Planer Y Y Y Y U U Y U Y U U U SAT

Tanner U Y U U U U Y U Y U Y N SAT

Diagnostic Test Appraisal

Study Clear study question

Appropriate reference standard

comparison

Diagnostic test,

reference standard applied

Reference standard

influenced test results

Disease status

described

Test methods described

Result presentation

Result applicability

Test applicability

All important outcomes considered

Impact of test use on population

Quality assessment

Tomonaga Y N Y N U U Y Y N Y U SAT

Qualitative Research Appraisal

Study Clear statement of aims

Qualitative methodology appropriate

Research design

appropriate

Recruitment strategy

appropriate

Data collection addressed

research aims

Researcher – participant relationship examined

Ethical issues taken into

consideration

Clear statement of

findings

Results valuable

Quality assessment

Law Y Y Y Y Y N Y Y Y SAT

Sodi Y Y Y U Y N N Y N ?

Tandjung Y Y Y Y Y N Y Y N SAT

CASP = Critical Appraisal Skills Program147;Y = yes; N= no; U= unclear; SAT = Satisfactory include in review; ?= Unclear if paper warrants inclusion in review

36

4.2 Objective 1: To describe the outcomes of patients who present in general practice with ACS symptoms and undergo cTn testing

4.2.1 Study characteristics

Five studies that addressed this review objective were included in the final analysis149-

153. The intervention examined in all studies was that of a cTn test requested by GPs,

with Aldous et al., Mann et al. and Tanner et al. examining laboratory based cTn testing

and Planer et al. and Tomonaga et al. assessing point of care cTn testing (POCT)

ordered by GPs in primary care.

One randomised controlled trial (RCT) was found (Tomonaga et al.) comparing POCT

to conventional diagnosis using best clinical practice, and the remainder were

observational cohort studies. While RCTs are preferable to observational studies when

evaluating interventions due to a lesser susceptibility to bias133, the RCT examined

POCT which is of less relevance to the review question. Hence the key messages of

the discussion of this section are drawn largely from findings of the observational

studies.

Across all studies, 3447 cTn test results were included in final analyses. The number of

participants overall is not known as this information was not provided by Mann et al.

4.2.2 Assessment of risk of bias within studies

Table 6 describes the assessment of risk of bias for the included studies.

4.2.3 Results of individual studies

The numbers of tests performed, results of tests and referral decisions in each study

are reported in Table 7.

All studies provided mortality data for patients with positive results, but only Aldous et

al. provided death rates for those with negative test results. The incidence of ACS and

other cardiovascular diagnoses were stated, with Aldous describing revascularisation

rates also. The two POCT studies also provided information on hospitalisation rates.

The period of follow up for outcomes varied from 5 weeks (Mann et al.) to 12 months

(Planer et al).

37

Table 7: Results of studies describing outcomes (Objective 1)

Study Number Hospital status – positive tests Hospital status – negative tests

Number of tests performed

Number positive

Number negative % positive Positive on

serial measure Negative test serial measures Hospital referral Hospital

admission Hospital referral Hospital admission

Aldous 2662 223 2439 8.4% 11/223 (4.9%) 297/2439 (12.2%) NR 184/223 (82.5%) NR 344/2439 (14%)

Mann 278 8 270 2.9% NR 12/270 (4.4%) 4/8 (50%) 4/8 (50%) 9/270 (3.3%) NR

Planer 349 5 344 1.4% NR NR 5/5 (100%) 5/5 (100%) 107/344 (31.1%) 42/344 (16%)

Tanner 11 11 0 100% NR NA 11/11 (100%) 11/11 (100%) NA NA

Tomonaga 147 19‡ 128 12.9% NR NR NR NR NR NR

Study ACS Delay Adverse events – positive tests Adverse events – negative tests Beneficial effects

Positive tests Negative tests Patients

with delay Median delay Died

Died within 48h

Died, admitted

Died, not admitted

Died Died within 48h

Died, admitted

Died, not admitted

Other NR

Aldous 101/223 (45.3%)

42/2439 (1.7%) NR NR 19/223

(8.5%) NR 17/223 (7.6%)

28/223 (12.8%)

34/2439 (1.10%) NR NR NR Arrhythmia 10%*

Heart failure 5%* NR

Mann 6/7 (86%)^ 12/270 (4.4%)^ 1 350

mins 0 0 0 0 0 0 0 NR Arrhythmia 1/270 (0.3%)

71/138 (51.4%)§

Planer 5/5 (100%) 18/344 (5.2%) NR NR 0 0 0 0 0 0 0 0 NR NR

Tanner 11/11 (100%) NA 11 485

mins 0 0 0 0 NA NA NA NA NR NR

Tomonaga 10/19 (52.6%) 7/128 (5.5%) NR NR 0 0 0 0 NR NR NR NR NR NR

* approximate values only; ^ GP likelihood of MI >10% ‡ 9 false positive results § 71/138 (51.4%) likelihood pretest >10%, post test <10%. Data only provided for 138/270 negative tests NR Not recorded NA Not applicable

mailto:19@

38

4.2.4 Synthesis of results

4.2.4.1 Deaths

Mortality rates in patients who had cTn testing were provided by Aldous et al., six

months after the test. The six month death rates were 8.5% in patients with positive

cTn tests and 1.1% for patients with negative test results. As the remaining studies

obtained complete follow up information by interview for patients with positive tests

after at least three days, it may be inferred that there were no early fatalities in the

other studies in patients with positive results.

4.2.4.2 ACS diagnoses

The rates of ACS in patients with positive cTn levels were over 50% in four of five

studies. The rate of ACS in the fifth study, by Mann et al., is not known with certainty,

with the rate of likely ACS reported as the GP’s estimation of likelihood of MI as the

cause for presentation.

4.2.4.3 Hospital admission, ACS complications and procedures

Across four studies, admission occurred in 50 -100% of patients with positive tests and

for 14-16% of patients with negative tests. Admission data were not provided by

Tomonaga et al. In the study by Aldous et al., not all patients with positive test results

were referred to hospital, with 17.5% being managed in the community. In regard to

ACS complications, Aldous et al., Mann et al. and Planer et al. reported incidences of

patients with heart failure or arrhythmias as discharge diagnoses, though exact figures

were not provided.

4.2.4.4 Delay

Delay was recorded by Tanner et al. and by Mann et al. In the study by Tanner et al.,

11 patients were sent for a cTn test prior to hospital. They were then referred to

hospital after a positive cTn result was returned, in a median time of 7.5 hours (range,

1 to 25 hours) after seeing the GP. The median time from symptom onset to hospital

arrival was 9.5 hours for this group, significantly longer than the 1.8 hours (range, zero

to 17 hours) for patients whose first health professional contact was with emergency

medical services.

Only one patient in the study by Mann et al. was thought to have had their admission

delayed by cTn testing, with a delay of five hours before admission.

4.2.4.5 Beneficial effects

There were potential beneficial effects of GP based cTn testing noted in the study by

Mann et al., with positive results returned in patients with late or atypical presentations.

39

4.3 Objective 2: To examine GPs’ understanding of cTn testing’s limitations, and the influence of cTn test result on management.

4.3.1 Study characteristics

A description of included studies is provided in Table 3. The studies by Aldous et al.

and Mann et al. were also assessed as part of Objective 1 of the review. Both used the

frequency of serial testing and timing of test ordering relative to symptom onset as

markers of GP understanding of limitations. Additionally Mann et al. used the difference

in GP assessment of likelihood of MI as a marker of management change.

Three new studies were examined as part of Objective 2. Law et al.154 used a postal

survey to all GPs listed on a university Department of General Practice database to

present nine hypothetical scenarios of MI likelihood, and to enquire whether a cTn test

would be ordered and whether the GP would wait for the result before referring patient.

The study by Sodi et al.155 consisted of a telephone survey of GPs who had ordered a

cTn test on a patient through one urban laboratory, to enquire about indications for the

test and proposed management before and after the test result. Tandjung et al.156 also

used a hypothetical scenario of a patient with a diagnosed ACS and a known cTn test

result to assess a GP’s action. The GPs’ responses were compared with those of a

group of cardiologists as a reference standard.

4.3.2 Assessment of risk of bias within studies

Table 6 describes the assessment of risk of bias for the included studies.

4.3.3 Results of individual studies

Table 8 reports the results of each study.

40

Table 8: Results of studies describing GP understanding and use of cTn (Objective 2)

Study GP understanding of test Effect on management

Sensitivity Specificity Indications

Number (%) of tests ordered during window period

Number (%) of pts with serial testing

Number (%) of patients with comorbidities causing false positive result

Number (%) ordered with inappropriate indication

Change in hospital referral decision

No change in hospital referral decision

Hospital referral despite negative result

Aldous (2012) NR 308/2439 (12.6%) NR* NR NR NR 14% of patients

Law (2006)

34% of GPs order with <5% risk MI 16% of GPs order with 5-50% risk MI 6% of GPs order with >50% risk MI

NR

Heart failure identified by 25% GPs PE identified by 25% GPs Renal failure identified by 39% GPs

NR

31-64% of GPs with <5% risk MI^ 10-78% of GPs with 5-50% risk MI^ 1-32% of GPs with >50% risk MI^

0-3% of GPs with <5% risk MI‡ 4-10% of GPs with 5-50% risk MI‡ 5-16% of GPs with >50% risk MI‡

NR

Mann (2006)

29 patients (12%) had tests performed; number of tests NR 3/29 (10.3%) NR NR 102/151 (67.6%)† 49/151 (32.4%) 9/270 (3.3%)

Sodi (2007) NR 0/16 (0%) NR

6/16 (37.5%) no indication provided 1/16 (6.25%) inappropriate (increased lipid level)

5/16 (31.25%) NR NR

Tandjung (2012) NR NR NR

175/471 (41.4%) inappropriate (to make diagnosis of known STEACS)

NR NR 6.7% of GPs

* Discharge diagnoses of admitted patients included heart failure, pulmonary embolism, arrhythmias ^ Percentage who would wait for result before referral; percent varies depending on duration ofsymptoms † Percentage with change in estimation of likelihood

‡ Percentage who would refer without waiting for result; percent varies depending on duration of symptoms

GP General practitioner NR Not recorded MI Myocardial infarction PE Pulmonary embolism STEACS ST elevation acute coronary syndrome

41

4.3.4 Synthesis of results

4.3.4.1 Knowledge of limitations

GP understanding of cTn’s limited sensitivity within the first 10 hours after symptom

onset was assessed directly by Law et al. and indirectly by Mann et al. and Aldous et

al.

Aldous et al. reported that 308 of 2439 (12.6%) had serial tests performed. The interval

of symptom onset to presentation was not reported so it is unclear how many patients

had an indication for serial tests. Notably, 11 of 223 (4.9%) patients with positive tests

had serial measures performed after elevation was detected on the first test. Law et al.

also posed hypothetical questions to respondents about the specificity of cTn testing.

Only a minority of GPs identified heart failure (25% of GPs), pulmonary embolism

(25% of GPs) and renal failure (39% of GPs) as causes of false positive results. One

scenario in the study by Tandjung et al. obtained useful information about GP

understanding of the indications for cTn, with 41.4% of GPs ordering a cTn test in a

patient with current chest pain and a STEMI on ECG, when cTn testing is not required

to make the diagnosis of ACS and would only delay referral to hospital.

4.3.4.2 Effect on management

The papers by Law et al., Mann et al., Sodi et al. and Tandjung et al. all collected

information about the effect of a cTn test result on GP estimation of likelihood of ACS

and on management of ACS.

In the scenario of a less than 5% likelihood of MI, most GPs in the study by Law et al.

would order the test and wait for the result before making a decision about referral.

In the Mann et al. group, 67.6% (102/151) of GPs would change their estimation of

likelihood of MI based on the test result. Three out of ten patients whose pre-test

clinical likelihood of MI was <10% moved to confirmed infarction after receipt of test

results, showing a definite change in estimation of likelihood of MI. Mann et al. also

reported that 86% of cTn tests were ordered on patients with a pre-test likelihood of MI

of less than 25%, and that pre-test likelihood fell substantially after the test result, with

84% of cases having a post-test likelihood of MI of less than 1%. In the scenario used

by Tandjung et al. of a patient with chest pain due to ACS that has been resolved for 8

hours, 93.8% of GPs would refer to hospital with a positive result, compared with 6.7%

with a negative result.

42

5. DISCUSSION

5.1 Summary of evidence

ACS and its complications were reported in all studies. While deaths occurred in

patients with positive cTn results, it was not clear if the deaths occurred early after

ACS, nor whether they were related to ACS. There were no early deaths in the studies

which specifically sought mortality data in the first few days following an ACS. This

provides some reassurance about the safety of test ordering. However, as Mann et al.

noted, “the overall number of positive tests was too small to conclude that management

of ACS was never adversely affected by cTn testing and waiting for results”.

Hospital admission occurred for over 50% of those with positive results and for 14 to

16% of those with negative results. Many non-admitted patients were over 75 years of

age, where a conservative medical or palliative approach due to age-related

comorbidities may have been more appropriate than admission and invasive

management. This could be considered a beneficial outcome if confirming ACS

diagnosis in a patient not suitable for hospital management allows progression to a

palliative stage of care. However, 2.6% of non-admitted patients underwent invasive

management with revascularisation within six months, suggesting that at least some of

the non-admitted patients may have been suitable for active management at their initial

presentation. Without knowing the residential and functional status of these patients, or

the timing of their symptoms, it is not possible to describe the late revascularisations as

either delayed or as adverse events.

A number of arrhythmias were described in the studies, including atrial fibrillation,

supraventricular tachycardia and complete heart block. Certain patients with these

arrhythmias may be at risk of adverse outcomes should the arrhythmias occur in the

community, as some treatments such as cardiac pacing are usually only available in

hospital. In order to classify these arrhythmias as adverse events they would need to

occur early while test results were awaited and while the patients were unmonitored in

the community, and this level of clinical information was not provided.

Similarly, heart failure was listed as an outcome for some patients. This is relevant in

that patients with heart failure manifesting as acute pulmonary oedema or cardiogenic

shock require urgent intervention not available in the community. Again, the limited

clinical information obtained in these studies prevents heart failure occurring early and

as an emergency being confirmed as an adverse outcome.

43

There was evidence of delay in diagnosis of ACS with some patients having their

admission deferred for some hours while awaiting a test result.

In patients with late or atypical presentations, both positive cTn test results and

diagnoses of ACS were seen, representing beneficial outcomes and therefore

appropriate test use. Due to the small numbers of patients in the included studies,

these benefits relate to individual patient cases rather than a systemic effect. There

was limited clinical information on the causes of these late positive results, and some

could have been due to conditions which may cause falsely positive results.

Nevertheless, some of the causes of false positive elevations, such as pulmonary

embolism or subarachnoid haemorrhage, carry a risk of death or major morbidity and

need intervention in their own right, and so their detection could be considered a

beneficial outcome, although cTn testing is not the preferred method to diagnose these

conditions.

In each study, some GPs showed evidence of inappropriate use of cTn testing. There

were instances of poor understanding of the limitations of cTn testing, with failure to

identify situations where false positive or false negative results were likely or where

there had been incomplete indications for ordering the test. It was evident that cTn

testing occurred within the 10 hour period of limited sensitivity, and certainly within the

24-72 hour period in which community based cTn testing is discouraged by major

guidelines129,131. Additionally, some patients with positive tests had serial measures

performed unnecessarily, as elevation was detected on the first test. This suggests a

need for education on the contraindications to serial testing as well as the indications. It

would be helpful to know if these tests were ordered by different providers, in which

case requesting of the second test may have been understandable if the first test result

was not available.

Many GPs in the studies would perform the test and wait for a result before acting in a

possible or probable MI scenario within 2 hours of symptom onset. As well as the risk

of being insufficiently sensitive, this delay raises the issue of safety for the unmonitored

patient. Other GPs in this context would order a cTn test but would send the patient to

hospital immediately without waiting for the results. This at least addresses patient

safety and reduces the chance of a missed diagnosis, even if the test is potentially an

inappropriate use of resources by not influencing management. There might be a

rationale for facilitating earlier diagnosis by hospital staff if the test is already being

44

processed while the patient is being transported to hospital, but this benefit is likely to

be lost given the time taken to collect and transport the specimen in the community.

CTn testing appeared to reduce the estimated likelihood of ACS in patients at low risk

of ACS, and to influence management in all four of the studies which collected

information from GPs. This would count as appropriate usage, though no measures of

statistical significance were provided.

45

5.2 Limitations

The studies within this review had a number of limitations and the above findings

should be interpreted with these in mind.

Selection and response bias affected a number of the papers. The study by Law et al.

was undertaken in collaboration with a University Department of General Practice. This

allowed the researchers to design the survey instrument in collaboration with GPs

affiliated with the department, and also to pilot the survey with a variety of GPs. The

high response rate of 72% shows the success of these measures. However, there were

acknowledged inaccuracies in the university database used to provide names of

participants. Furthermore, the design of this study carried the risk of response bias; the

identifiability of respondents to authors who were known to them as GP colleagues

increased the likelihood of the reporting of desired outcomes. The use of non-

identifiable data may have overcome this issue.

The nature of the cTn assay used was not described for the POCT studies. There was

initial training in the use of the POCT kit but no quality assurance process for the GPs’

ability to perform or interpret the test kit result. This is relevant as the majority of false

negative results in the study by Tomonaga et al. came from the same practice, with the

authors suspecting misuse of the device. The exact method of cTn testing for all

patients was not stated by Law et al. or Tandjung et al., which affects the

generalisation of their findings to the urban laboratory based GP testing. Though not

specifically mentioned as POCT, Tandjung et al.’s scenario involved “perform(ing) a

troponin test in my practice”. POCT is known to be widely used in northern Europe and

76.3% of respondents in this study had POCT for cTn as part of their practice

infrastructure. Nineteen GPs in the study by Tandjung et al. stated that they would

perform cTn testing despite not having this available in their practice, implying use of

laboratory based cTn in these cases. Similarly, in the scenario by Law et al. where

there was a greater than 50% likelihood of MI, 68 GPs (32%) of GPs would wait for a

cTn result, which was raised as a concern by the authors. It is known that five GPs in

this group came from rural areas where POCT may have been used. The rapid result

obtained by POCT means that waiting a shortened time for a result would be an

appropriate management option in this situation.

The accuracy of assessment of outcomes was low in all laboratory studies.

Questionnaires to GPs about their clinical practice may be unreliable and this was a

risk in the three studies using survey data, with no independent verification of

46

questionnaire responses despite key conclusions being based on this information. The

main study addressing delay, by Tanner et al., failed to obtain objective measurements

of time intervals beyond those recalled by the patients, nor was there corroboration of

presenting symptoms as reported by the patient to the GP. These authors concluded

that cTn testing by GPs represented a failure to understand the risks of delay in

diagnosis and treatment of suspected ACS. However, if the presenting symptoms

described to the GP were more benign than those recalled by the patient after hospital

admission, the GP’s decision not to refer may have been appropriate. Additionally,

while the questionnaire used by Mann et al. was sent to GPs within 3 days of ordering

of the test, there is still the possibility of recall bias in the case of incomplete medical

records. There were missing data from the group of patients with negative results in

Mann et al. which affects the number of patients with a potential beneficial outcome.

The use of hospital databases would have improved the accuracy of assessment of

outcomes, but these were underutilitised. Instead, Mann et al. relied on the

questionnaire response of GP estimate of MI likelihood as a surrogate marker for

diagnosis in most patients. When hospital databases were used in the laboratory

studies to verify GP responses, the databases and items used to arrive at a diagnosis

of ACS were not specified. Even with the use of health record databases, actual

information on delays in inpatient therapy may only be obtained by detailed review of

hospital medical notes, and there are significant privacy considerations in obtaining this

clinical information which may have deterred the authors from proceeding with such

research.

The accuracy of outcome measurement in the POCT studies was variable. A major

weakness of the study by Tomonaga et al. was the lack of reliability of the reference

standard of conventional best practice diagnosis. It was unclear if guidelines or other

education were used to direct best practice or if diagnosis was left to the individual

doctor's clinical judgement. In this study there was a lack of independence of the index

test and reference standard and a failure to involve an independent blinded assessor,

with the same clinician performing the clinical diagnosis, cTn test and the final

diagnosis. It is likely that these clinicians incorporated additional information into the

final diagnosis 3 weeks after the test, as specialist reports and hospital data were

provided to GPs on patient discharge for those admitted to hospital. Practical and data

protection reasons were cited as the reasons for this aspect of the study design.

47

The paucity of clinical information reduces the ability to draw conclusions for either

objective of the review. Regarding Objective 1, a clinical context for each test would

have assisted in classifying outcomes as adverse or beneficial. Additionally, clinical

and demographic information would also allow assessment for confounding, which was

not fully identified or addressed in any study. Non-coronary comorbidities causing false

positive results could have accounted for any or all of the positive tests. Regarding

Objective 2, many GPs found it difficult to indicate management based on cTn testing

alone, stating that clinical factors would influence their decision making and risk

stratification. Removing this clinical context from decision making does not mirror

actual GP practice, limiting the strength of study findings. Tandjung et al. did supply

GPs with this information, making the findings in that study more representative of real

world practice.

Additionally, no study recorded the patient’s contribution to the decision to have cTn

testing. In the study by Tanner et al., 27 of 47 patients who chose to see a GP did not

believe their condition was serious and six of 47 believed the hospital system was too

busy to provide their care. It may have been that this reluctance on the part of the

patient to attend hospital was the driver for the GP to find evidence of current ACS and

to use this evidence to convince the patient that hospital attendance was necessary.

Previous studies have used carer surveys, hospital records157, data from the referring

GP138 and benchmarking processes103 to improve the accuracy of recorded times and

the inclusion of these processes in the design of the study by Tanner et al. would add

weight to the authors’ conclusions.

The review itself is limited by the small number of eligible studies, as well as the

marked heterogeneity of patient populations and outcome measures in studies that

were the included. Participant numbers were low in all studies and most reported data

from single centres. Publication bias might account for this paucity of data. Laboratory

audits and audits as part of GP quality improvement cycles are often performed for the

purposes of internal review, and evidence of good practice, while reassuring for

individual clinicians and laboratories, may not merit wider dissemination.

48

5.3 Implications for future research

The optimal study method to address this review’s objectives would firstly address the

risk of bias by using non-identifiable data from a group of GPs who are representative

of the wider GP population. In terms of the intervention used, the involvement of

multiple pathology providers would strengthen findings of further studies. Results from

the studies included in this review can only be generalised to the wider population if the

population serviced by that laboratory is representative of the region, and only Mann et

al. claimed to have a representative cohort. The ideal study would also use laboratory

testing rather than POCT in order to be to be generalisable to urban Australian GPs, as

this is the common method of cTn testing in urban primary care in this country.

Obtaining clinical context for the patient undergoing testing would be crucial to

classification of events subsequent to the ACS as adverse or beneficial outcomes.

Clinical information would also contribute to the understanding of the incidence of false

positive and negative results and to the rationale for the GPs’ referral decisions.

Deficiencies in outcome measurement could be addressed by making wider use of

linked data sources to provide objective measures of hospital events and assist with

the reporting of outcomes, for a defined period of follow up. The results of outpatient

investigations confirming a diagnosis of ACS, such as positive dynamic testing that was

not confirmed with inpatient angiography, would need to be captured to obtain a true

picture of ACS events. However this would be a challenge in Australia given the

abundance of private providers of cardiology investigations.

The preferred method of investigating adverse outcomes would be a randomised

controlled trial designed to detect rates of adverse outcomes in patients who did and

did not undergo testing. The observational nature of the three laboratory based studies

in this review means that a comparison of outcomes with patients who did and did not

undergo GP investigation was not possible. However randomisation of a patient with

symptoms suggestive of ACS to community care would be unethical and unacceptable

to both patient and doctor. Additionally, the heterogeneous nature of ACS

presentations in general practice poses a risk of bias and if the low absolute numbers

of positive tests in the studies included in this review are a guide, a fully powered study

would need to be very large and therefore probably unfeasible. An observational

design remains the realistic method to evaluate patient outcomes.

49

6. CONCLUSION

In summary, knowledge about the appropriateness of GP initiated cTn testing is limited

due to a small number of published studies, the heterogeneity of these studies and the

small number of patients in existing studies. The understanding of GP cTn test use

would be improved if there were clinical data on patients in the existing studies, as well

as precise and complete outcome data.

What is evident from this review is that information on the outcomes of GP cTn testing

and the effect of cTn testing on management needs to be interpreted within a clinical

context. While it appears from this review that cTn testing occurs in all clinical

scenarios regardless of likelihood of pain and recency of symptoms, the majority of

these data comes from hypothetical scenarios, and the findings based on actual GP

practice involve such small cohorts that their validity is limited.

This provides justification for the research presented in this thesis evaluating cTn

testing by GPs, where non-identifiable clinical information has been provided by GPs

on patient presentation and risk status, and linked with outcome data from hospital

record databases. This will account for some of the limitations in the literature detected

by this review and should further the understanding of what is appropriate use of cTn

testing by GPs.

50

CHAPTER THREE: METHODS 1. STUDY DESIGN

This study was a prospective cohort design.

2. SETTING

The study recruited patients who had cTn blood tests ordered by a general practitioner

in a non-hospital setting and who had their sample collected at community collection

centres of two laboratories in urban Perth, Western Australia. The laboratories were

PathWest and St John of God Pathology, two of the five pathology laboratories in

Perth. Together, these laboratories cover all Perth regions. The period of recruitment

was 24 September 2009 to 3 September 2010, with recruitment ceasing so that follow

up could be completed within the four year period specified by UWA for the completion

of the research.

Ethical approval to conduct the survey was obtained from Human Research Ethics

Committee at the University of Western Australia (RA/4/1/2275; 13 July 2009), St John

of God Hospital (370; 7 May 2009), the Department of Health Western Australia

(2013.04.02; 9 April 2013) and the South Metropolitan Health Services Board (08.136;

28 August 2014). The Medical Directors of St John of God Pathology and PathWest

gave written consent for provision of laboratory data.

3. PARTICIPANTS

3.1 GP cohort

Consecutive patients with samples collected at collection centres of the two

laboratories were included in this study. Both laboratories used the standard sensitivity

cTnI assay. A Research Assistant employed by the laboratories then approached

requesting GPs to complete a survey which described the clinical scenario leading to

the test request, and the clinical course of the patient following the test result.

3.2 ED cohort

The Multiple Infarct Markers in Chest Pain (MIMiC) study dataset158 was used to

compare clinical presentations and outcomes in the GP survey cohort with an

Emergency department cohort. This prospective cohort study was conducted between

September 2008 and June 2009 in two tertiary referral hospitals and three general

hospitals in urban Perth. The urban catchment areas of these five hospitals were

similar to those of the collection centres used in the GP survey phase.

51

Participants were a representative sample of patients undergoing evaluation for

possible ACS with serial cTn testing. Exclusions were patients aged less than 18 years,

those who were pregnant or those where ECG criteria called for urgent reperfusion

therapy, such as patients with ST segment elevation ACS on their initial ECG.

3.3 Linked data

The Department of Health Western Australia Data Linkage System (WADLS) uses

computerised probabilistic matching supported by clerical officers to connect all

available health and related information for the WA population159. Connections, or

linkages, are created by comparing the personal information available and calculating

the likelihood that records belong to the same person, place or event, while at the

same time protecting personal privacy160.

Linked data were obtained for all patients for a minimum twelve month period following

the date of the test, regardless of whether the GP responded to the survey. The final

cTn test included in the study was performed in September 2010, and follow-up

continued until October 2011. The Emergency Department Data Collection, the

Hospital Morbidity Data Collection and the Mortality Register Data Collection were

searched for the period from 24 September 2009 to 3 October 2011. Linkage and

extraction were performed in November 2013 to compensate for delay in updating of

Department of Health records.

Patients who had samples collected at rural and regional centres were excluded, since

it was thought that GPs in these centres may use cTn testing differently to urban GPs.

Reduced access to tertiary hospital EDs in rural and regional centres may result in the

need to obtain greater evidence of ACS before instigating transfer of patients to larger

centres. Additionally, point of care assays may substitute for laboratory based testing in

rural areas, and so the prevalence of cTn testing may have been underestimated if

laboratory results alone were examined. If the collection centre location was unknown,

as occurred when the specimen was collected within doctor’s rooms or by domiciliary

collection, the patient’s postcode was used for allocation to rural, regional or urban

areas.

52

4. VARIABLES

4.1 GP cohort

Tables 9 and 10 list the data variables obtained from each source in this phase. A copy

of the questionnaire used in the GP survey phase is provided as an appendix.

Table 9: Data variables obtained by laboratories

Variables obtained from GPs Variables obtained from laboratory records GP name and contact details Time from sample collection to registration Patient name, address, gender Time from registration to result availability Patient date of birth cTn result (quantitative) Collection centre ID code Interpretation of cTn result Collection centre postcode Renal function where available (Cr or eGFR) Date of test request Date and time of sample collection

Cr = plasma creatinine eGFR = estimated glomerular filtration rate

Table 10: Data variables obtained by laboratories from requesting GP

Patient’s symptoms prompting ordering of cTn test Duration of symptoms prior to test ordering Nature of any pain

Typical of cardiac ischemia Atypical of cardiac ischemia

Non-pain symptoms that may represent cardiac ischemia Dyspnoea Syncope Dizziness Palpitations Fatigue

Cardiovascular risk factors known to be present at the time of presentation Current smoker or previous smoker of more than10 pack years Hyperlipidaemia Hypertension Diabetes Personal history of CHD or equivalent Family history of CHD (1st or 2nd degree relative less than 60years old at onset)

Assessment by the GP of the likelihood of ACS before and after the test* Less than 5% 5-10% Greater than 10%

Intended management before and after receipt of test result Referral to Emergency Department. Cardiology review as an outpatient Ongoing management by GP without referral

CHD= coronary heart disease; ACS = acute coronary syndrome; * = category scores chosen to be the same as those used in two of four articles in the systematic review which examined GPs’ estimation of likelihood149, 154

53

4.2 ED cohort

Variables obtained from the ED cohort data set are listed in Table 11.

4.3 Linked data

Table 12 lists specific outcomes provided by the Data Linkage System.

Table 11: Data variables obtained from the ED cohort dataset

Age Gender Cardiovascular risk factors

Current smoker Hyperlipidaemia Hypertension Diabetes Personal or family history of coronary heart disease or equivalent

Presenting symptom Admission status Outcomes

All-cause mortality MI Unplanned revascularisation

MI = myocardial infarction

Table 12: Data variables from Data Linkage System

Emergency Department Data Collection Hospital type Presentation date and time (DDMMYYYY) Transport mode (arrival) Referral Source Triage category Presenting problem Principal diagnosis Departure destination

Hospital Morbidity Data Collection Admission age Gender Hospital category Admission date Separation date Length of stay Days in Intensive Care Unit Hours on continuous ventilatory support Mode of separation (method and destination of patient discharge) Principal diagnosis Co-diagnosis Additional diagnoses Principal procedure Additional procedures

WA Mortality Register Date of death Cause of death

54

4.4 Outcomes

Information on hospital presenting problems, diagnoses and procedures are listed in

Tables 13 to 18. This information was obtained from the Emergency Department Data

Collection (EDDC) dataset and the Hospital Morbidity Data Collection (HMDC).

Outcomes were defined according to standardised data definitions recommended for

use in Australasian ACS research137. Specific diagnosis and procedure codes

representing these outcomes were selected from the International Statistical

Classification of Diseases and Related Health Problems, Tenth Revision, Australian

Modification, 6th and 7th editions161,162, and the Australian Classification of Health

Interventions 7th edition163.

Significant clinical events representing adverse outcomes of ACS, as listed in Table 18,

were those occurring within a 30 day time period following ACS, as this is the interval

where there is the greatest risk of death, progression to MI or the development of

recurrent MI. After 30 days, most patients with ACS resume a clinical course similar to

that of patients with stable coronary disease9,138.

If a specific diagnosis was recorded that excluded a cardiovascular cause for a

presenting problem or diagnosis, such as chest pain with a coexistent diagnosis of

chest wall trauma or dyspnoea with a coexistent diagnosis of asthma, that patient’s

data was reclassified as non-cardiovascular.

Table 13: EDDC cardiovascular presenting problems

Problem group Specific symptom Chest pain

Presence of acute chest, epigastric, neck, jaw or arm pain or discomfort or pressure without apparent non-cardiac source

Associated symptoms

Symptoms that developed since pain/discomfort started, and cannot otherwise be explained by a non -cardiac condition, including: Nausea, vomiting, Sweating, clamminess Syncope, blackout, unexplained loss of consciousness Palpitations, arrhythmia Shortness of breath, dyspnoea, breathlessness

General/atypical symptoms Fatigue, nausea, vomiting, diaphoresis, faintness, back pain

Symptom recorded as provisional diagnosis

Cardiac arrest Acute myocardial infarction, unstable angina Heart failure, pulmonary oedema Hypotensive, hypertensive Arrhythmia Cardiomyopathy

55

Table 14: EDDC cardiovascular presenting problems – ICD-10AM codes

Symptom group ICD-10AM code Chest pain QG000 QGA00 QGB00 QGC00 Associated symptoms HJ000 HQ000 HQA00 KF000 KN000 BE000 BEA00 BEB00 BEC00 CF000 CFA00 General/atypical symptoms BD000 CJ000 Cardiac arrest BA000 BAA00 BAB00 Acute myocardial infarction SDB00 SXB00 Unstable angina SDV00 SXP00 Heart failure, pulmonary oedema SDM00 SXG00 Hypotensive, hypertensive BB000 BC000 BD000 Arrhythmia SDC00 SDCA0 SDD00 SDE00 SDG00 SDU00 SDW00 SXC00 SXCA0 SXCB0 SXCC0 SXCD0 SXCE0 SXCF0

Table 15: EDDC and HMDC principal and additional diagnoses

Diagnosis group Specific diagnosis ACS STEMI, NSTEMI, AMI, unstable angina CHD CHD, Angina NOS, atherosclerosis of autologous bypass graft

Atherosclerosis

TIA Hypertensive heart disease Intracerebral haemorrhage Cerebral infarction AAA rupture Acute vascular disorder of intestine

Arrhythmia

VT, SVT, Atrial fibrillation Atrioventricular block Bradycardia Cardiac arrhythmia - other

Other CVS diagnosis

Cardiogenic shock Symptomatic aortic stenosis Cardiomyopathy Heart failure Syncope Mechanical complication of cardiac device Other breathing abnormality Dizziness Palpitations Chest pain - anterior chest wall Chest pain on breathing Chest pain unspecified

Precordial pain ACS = acute coronary syndrome, STEMI = ST elevation myocardial infarction, NSTEMI = Non-ST elevation myocardial infarction, AMI= acute myocardial infarction, CHD = coronary heart disease; NOS = not otherwise specified, VT = ventricular tachycardia, SVT = supraventricular tachycardia, AAA = abdominal aortic aneurysm, TIA= transient ischaemic attack

56

Table 16: EDDC and HMDC principal and additional diagnoses - ICD-10AM codes

Diagnosis group ICD-10AM code ACS I20.0 I21 I22 I23 CHD I20.1 I20.8 I20.9 I24 I25 Arrhythmia I44 I45 I47 I48 I49 Atherosclerosis G45 I11 I61 I63 I70 I71 I72 I73.8 I73.9 I74 K55 Other CVS diagnosis I30 I31 I32 I33 I50 I51 I52 I95.1 J81 R01 R01 R02 R03 R42 R50 R53 R55 R57.0

ACS = acute coronary syndrome, CHD = coronary heart disease, CVS = cardiovascular

Table 17: ACHI codes for procedures

Outcome ACHI code

Emergency or urgent revascularization procedure*

30500-04 30500-05 30306-00 35303-01 38215-00 38218-00 38218-01 38218-02 38300-00 38300-01 38303-00 38306-01 38306-02 38306-03 38306-04 38306 -05 38497-00 38497-01 38497-02 38497-03 38497-04 38497-05 38497-06 38497-07 38500-00 38500-01 38500-02 38503-00 38503-01 38503-02 38503-03 38503-04 90201-00 90201-01 90201-02 90201-03 90201-00

*ACHI codes apply to all revascularisation procedures irrespective of urgency

Table 18: Adverse outcomes and ICD-10AM codes

Outcome ICD-10AM code Death from cardiovascular cause I21 I22 I23 I24 I25 Death of uncertain cause R96 R98 R99 Cardiac arrest I46 Cardiogenic shock R57.0 Unstable angina I20 Acute myocardial infarction I21 I22 I23 I24 I25 Ventricular arrhythmia I47.2 High degree AV block I44.3 I44.3 Heart failure requiring intervention I50

AV= atrioventricular 5. MEASUREMENT

5.1 GP cohort

Information was collected using a one page survey sent and returned by facsimile, with

an information sheet and consent form sent at the same time. GPs were contacted

within one week of testing, with telephone follow up to non-responders performed one

week after initial contact. GPs who had previously indicated that they did not wish to be

involved with the study were not contacted in regard to subsequent tests that they

requested.

All identifying information for patients and GPs, including patient name, address,

gender and date of birth was removed from laboratory data before the data were

57

provided to the researchers. Each test result was given a unique project number, being

the laboratory’s sample identification number.

The use of a unique project number for each sample, rather than a number for each

patient, meant that additional methods were required to identify patients who had more

than one test performed. The unique identifier assigned by the WADLS as part of the

linkage process described below was used for this purpose.

5.2 ED cohort

Clinical data had been obtained by the treating clinician in ED at the time of

presentation. Outcome information had been sought by a research nurse who made

telephone contact with patients, relatives, GPs or cardiologists, supplementing from

these sources with the hospital discharge summary as necessary.

Non-identifiable data were provided to the researchers of this study by the MIMiC study

investigators via encrypted Excel documents.

5.3 Linked data

Identifying information from the laboratories was retained by Research Assistants

working within the laboratories and provided to the WADLS, along with a unique project

number. Following linkage and extraction, identifying data variables were removed by

the WADLS before delivery to the researcher, who used the unique project number to

merge the non-identifiable data from the laboratories with the non-identifiable data from

the WADLS.

The unique identifier assigned by the WADLS as part of the linkage process was used

to identify patients who had more than one test performed. Records were excluded

from analysis if there was no principal diagnosis stated. Records were also excluded if

the presenting symptom or principal diagnosis was insufficiently specific to allow

classification as a possible ACS symptom (for example, pain or respiratory complaint

without additional information). Duplicate records with more than one hospital

admission for the same patient on the same day were considered as one admission for

the purposes of statistical analysis, for example inter-hospital transfer recorded as a

new admission.

All data transfers between laboratories, the WADLS and the researchers used Winzip

256bit AES encryption.

58

6. STUDY SIZE

A retrospective review of GP-initiated cTn tests was performed by the two pathology

providers for the purposes of estimating sample size. It showed that over a five month

period in 2008, 1,596 cTn tests were requested by GPs, at a rate of approximately 45

tests per week, 55 of these tests (3.4%) being positive results. Urban GPs requested

approximately 35 urban tests per week with a 5% rate of positive tests. On this basis a

sample size of 1400 tests was expected with 70 tests expected to be positive.

7. QUANTITATIVE VARIABLES

Outcomes were grouped into adverse or beneficial as shown in Table 19.

Late presentation was defined as greater than 48 hours after symptom onset, 48 hours

being the time period following ACS where inpatient continuous cardiac monitoring is

indicated7,27,54.

Atypical presentations included atypical pain and the presence of non-pain symptoms

in the absence of pain. Atypical pain was defined as back pain, sharp pain or pleuritic

pain. Non-pain symptoms included dyspnoea, palpitations, nausea, vomiting,

sweating, fatigue, syncope or presyncope137.

Table 19: Grouping of quantitative variables

Outcome Event Adverse Death from cardiovascular cause Death from uncertain cause Cardiac arrest Cardiogenic shock AMI - revascularisation AMI – no revascularisation Unstable angina – revascularisation Unstable angina – no revascularisation Ventricular arrhythmia High degree AV block Heart failure requiring intervention

STEMI and delay greater than 120 minutes from time of presentation to GP to revascularisation, if performed.

STEMI and delay greater than 12 hours from onset of symptoms to revascularisation, if performed.

Beneficial Late or atypical presentation plus any of: AMI or unstable angina Emergency revascularization procedure Urgent revascularization procedure Elective revascularization procedure

AMI = acute myocardial infarction, AV = atrioventricular, STEMI = ST elevation myocardial infarction,

59

8. STATISTICAL METHODS

All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC,

USA)164 and XLSTAT-Pro (Version 2014.2, Addinsoft Inc., Brooklyn, NY, USA165.)

Categorical outcomes were summarized using frequency distributions. Tests of

normality were applied to the continuous outcomes using Anderson-Darling test and P-

P graphs. For normally distributed variables, the differences between group

characteristics were assessed using the two sample t test for continuous variables and

the chi squared test and Fisher’s exact test where appropriate, based on expected

frequencies for dichotomous variables.. A p value of <0.05 was considered statistically

significant.

60

CHAPTER FOUR: RESULTS 1. PARTICIPANTS

Figure 5 describes the flow of participants through the phases of the study. 369

surveys were sent to 127 GPs. 124 surveys, each on a unique patient, were completed

by 57 GPs, representing a response rate of 38%. There were no significant differences

between included and excluded patients in respect of age or gender (each P >0.10).

Two GPs declined to participate after receiving the first survey and requested that no

further surveys be sent to them, with no reason provided for their decision. These two

GPs requested 41 tests on 39 patients.

Figure 5: Flow chart of participants

Troponin tests with linked data n=361

Direct admissionn=22

<48 hoursn=17

1 or more hospital presentations during follow upn=94

ED presentationn=72

48 hours to 30 days

n=15

>30 days n=40

ACSn=6

No ACSn=11

ACSn=4

No ACSn=11

ACSn=10

No ACSn=30

<48 hoursn=3

48 hours to 30 days

n=6

>30 days n=13

ACSn=0

No ACSn=3

ACSn=2

No ACSn=4

ACSn=5

No ACSn=8

61

2. DESCRIPTIVE DATA

2.1 Characteristics of GP cohort with survey data

Table 20 presents characteristics of the 124 patients in the GP cohort for whom survey

data were available. The most common presentation was pain typical of cardiac

ischemia, with 56% of patients reporting this symptom. Twenty two per cent had pain

atypical of ACS and 19% had no pain, instead presenting with dyspnoea, syncope,

dizziness, palpitations or fatigue. Two per cent had no symptoms of ACS prior to cTn

testing. Reasons given for cTn testing in these patients were: monitoring of cardiac

side effects in one patient on antipsychotic medication, monitoring of cardiac symptoms

in one patient on lipid-lowering therapy, and monitoring for emergence of ACS in one

patient with ongoing creatine kinase elevation.

Sixty-two patients (51%) had the onset of symptoms prompting the test request within

48 hours of test requesting, and 29 patients (23%) had symptoms within 12 hours of

the request for the cTn test.

Data concerning CHD risk factors were available for 104 patients. Two or more risk

factors were present in 38% of patients, placing them at increased risk of ACS,

according to the NHF/CSANZ risk stratification framework8,27,166. Nineteen per cent

were automatically at high risk of ACS, including 7% with a personal history of CHD,

6% with typical symptoms and diabetes and a further 7% aged over 60 years with

diabetes. Of the 85 patients who had renal function recorded by the laboratory, 26%

had some degree of renal impairment with an eGFR of less than 60mls/min/1.73m2,

another factor that elevates risk8.

Twenty-one patients did not have complete data for all CHD risk factors; 20 St John of

God (SJOG) patient surveys did not provide information on the presence of a personal

history or family history of CHD.

2.2 Comparison with Emergency Department cohort

A comparison with the ED cohort of 1758 patients showed significant differences in the

prevalence of all individual CHD risk factors with the exception of diabetes, with greater

proportions for all risk factors in the ED cohort (Table 21).

There was a significant difference in the proportions of participants with no known risk

factors (P <0.01), one or more risk factors (P <0.01) and two or more risk factors (P

<0.01).

62

Table 20: Characteristics of included patients

GP (n=124) ED (n=1758) n % n % P

Median age (IQR) 61 (45-73) 62 (50-74) 0.38 Male 55 44 984 56 <0.01 Test result Positive 2 1.6 168 10.7 <0.01

Negative 122 98.4 1408 89.3 <0.01 Risk factors* Smoker 15 12 425 24 0.01

Hypertension 51 41 923 53 0.02 Dyslipidaemia 47 38 842 48 0.03 Diabetes 15 12 327 19 0.07 Past Hx CHD 8 6 621 35 <0.01 Family Hx CHD 24 19 879 50 <0.01

Presenting symptoms

Typical pain 69 56 Atypical pain 27 22

Non-pain symptoms

24 19

No symptoms 3 2 NR 1 1

Pain duration Less than 12h 29 23 12-48h 33 28 More than 48h 57 46 NR 5 2

eGFR <30 3 2 30-60 19 15 >60 63 5 NR 39 31

*104 patients with complete risk factor data; **1576 patients with 8-12 hour cTn levels; Hx = history, CHD = coronary heart disease; NR = not recorded

Table 21: Comparison of risk factors between GP and ED cohorts

GP (n=104) ED (n= 1758) Number of risk factors n proportion (CI) n proportion (CI) P Zero 25 0.24 (0.17, 0.33) 167 0.09 (0.08, 0.11) <0.01 One 40 0.38 (0.30, 0.48) 408 0.23 (0.21, 0.25) 0.01 Two 19 0.18 (0.12, 0.27) 445 0.25 (0.23, 0.27) 0.10 Three 17 0.16 (0.10, 0.25) 373 0.21 (0.19, 0.23) 0.20 Four 3 0.03 (0.01, 0.08) 365 0.21 (0.19, 0.23) <0.01

2.3 Patients presenting within 12 hours of symptom onset

Data for a subgroup of 29 patients of the GP cohort who presented within 12 hours of

symptom onset is shown in Table 22. The majority (69%) of these patients had pain

typical of cardiac ischemia. Twenty four patients (83%) had at least one CHD risk

factor including five patients (19%) with a personal history of CHD. Figure 6 shows

patients who had a combination of these characteristics. Fifteen patients (52%) had

typical pain as well as at least one CHD risk factor.

Ten patients in this subgroup of 29 (34%) had accurate information on the timing of

sample collection, with nine patients (34%) having specimens collected less than one

63

hour after presentation. This represents a maximum 13 hour interval between symptom

onset and specimen collection. The median time taken for result provision for this

group was 121 minutes (range, 23-1466).

Table 22: Characteristics of GP patients presenting within 12 hours

Total (n=29) % Presenting symptoms Typical pain 20 69

Atypical pain 6 21 Any pain 26 90 No pain 3 10 No symptoms 0 0 NR 0 0

Risk factors Smoker 3 10 Hypertension 10 34 Dyslipidaemia 15 52 Diabetes 3 10 Family history CHD 8 28 Personal history CHD 5 17

Cumulative risk factors Zero 5 17 One 13 45 Two 6 21 Three 5 17 Four 0 0

CHD = coronary heart disease; NR = not recorded

Figure 6: Number of patients with typical ACS pain, symptom onset <12hours and CHD risk factors

Symptom onset <12 hours prior to presentation

At least one CHD risk factorTypical pain

15

5

26

5

4

3323

64

2.4 Test result availability

The time taken for result availability is shown in Table 23. The total time from specimen

collection to a result being available was divided into two intervals. First, the time from

specimen collection to registration represented the time taken for transport of the

specimen from a specimen collection centre to a processing laboratory, including any

delay at the collection centre awaiting specimen being collected by the laboratory’s

courier. This interval varied depending on the location of the specimen collection

centre, with collection centres that were co-located with laboratories having the

shortest time interval. The median time for specimen collection to registration was 31

minutes (range, 0-1465).

Second, the time from specimen registration to result availability represented the

processing time for the specimen. The median time for this interval was 59 minutes

(range, 1 - 391). Overall the total median time from specimen collection to result

availability was 128 minutes (range, 23 -1466).

Table 23: Time in minutes from specimen collection to result availability

Interval Median (IQR) Pathwest SJOG Specimen collection to result availability 128 (78-190) 93(50-186) 135 (94-192)

Collection to registration 31 (12-88) 11 (6-29) 51 (21-113) Registration to availability 59 (30-111) 77(41-148) 52 (27-101)

65

2.5 Effect of test result on GPs’ estimation of likelihood of ACS

One hundred and twenty GPs provided a response to this item. Results are shown in

Table 24. The majority of GPs (80, or 67%) estimated the likelihood of ACS to be low,

or less than 5%, prior to receipt of test results. This proportion increased to 110 GPs

(92%) after test results were received. Of the two patients with positive results, one GP

estimated the likelihood as intermediate before the test, the other as high, and both

GPs estimated the likelihood as high after receipt of test results.

For the 39 patients with an estimated intermediate or high likelihood, seven (18%) had

symptom onset within 12 hours of presentation, and 20 (51%) had symptom onset

within 48 hours of presentation.

A significant proportion of GPs (27.5%) changed their assessment of the likelihood of

ACS, in response to a negative test result (chi-square test, P <0.01). No further

information on the clinical circumstances was available where the estimation of

likelihood of ACS increased following a negative test result.

Table 24: Effect of test result on GPs’ estimation of likelihood of ACS

Estimated likelihood No. (%) before test No. (%) after test P Low (<5%) 80 (67) 110 (92) <0.01 Intermediate (5-10%) 31 (26) 6 (5) <0.01 High (>10%) 9 (7) 4 (3) 0.25

2.6 Effect of test result on GPs’ intended management

One hundred and eighteen GPs responded to this question, as shown in Table 25.

Most GPs (85, or 72%) intended to manage the patient themselves before the test

result. This increased to 97 GPs (82%) after the test results were known.

Despite the test result having a significant effect on estimated likelihood of ACS, the

test result did not significantly influence a GP’s intended management, with 22.9% of

GPs changing their intended management (chi-square test, P =0.23).

Table 25: Effect of test result on GPs’ intended management

Intended management No. (%) before test No. (%) after test P GP 85 (72) 95 (82) 0.17 Outpatient cardiology 15 (13) 11 (9) 0.53 ED 18 (15) 10 (8) 0.16

66

3. OUTCOME DATA

3.1 Emergency Department cardiovascular presentations

Linked data were available for 361 tests performed on 355 patients, with eight patients

unable to be linked due to insufficient identifying information from the requesting GP.

Table 26 shows ED cardiovascular (CVS) presentations for this group. There were 112

presentations to ED with CVS symptoms by 76 patients (21.4% of those with linked

data available) during follow up. Symptoms were: chest pain (84 patients), dyspnoea

(15), palpitations (8), dizziness (7) and syncope (5).

112 presentations had an ED diagnosis of a CVS condition, of which 20 presentations

in 16 patients were due to ACS and 92 presentations were for other conditions, as

described in Table 27. Eighty-seven presentations were given a triage category of 1 or

2, requiring immediate medical review or review within 10 minutes. 20 of these high

triage category presentations had an eventual diagnosis of ACS.

Within 30 days of cTn testing, 24 patients (21.4%) who presented to ED with a CVS

symptom were given a CVS diagnosis, of whom nine (8%) had an diagnosis of ACS

and 15 (13%) had other CVS diagnoses (two with 2nd degree atrioventricular block, one

with heart failure, one with atrial fibrillation and the remainder with unspecified chest

pain.)

By comparison, the ED cohort included 357 patients (20.3%) with an ED diagnosis of

definite ACS, and 984 (56.0%) with an ED diagnosis of definite or possible ACS.

Table 26: ED presentations – GP cohort (n=355)

Diagnosis n (%) Number of ED presentations ACS

Non-ACS 20 (17.9) 92 (82.1)

Total 112 (100) Number of patients presenting to ED ACS

Non-ACS Total

16 (14.3) 60 (53.6) 76 (67.9)

Age of patients presenting (median, IQR) ACS Non-ACS Total

64 (55,80) 60 (46,77) 66 (52,78)

ED presentations with CVS diagnosis within 1 month of test ACS 9 (8.0) Non-ACS 15 (13.4) Total 24 (21.4)

ED presentations requiring admission ACS* 19 Non-ACS 38 Total 57

*One patient with UA was not admitted; ACS = acute coronary syndrome; CVS = cardiovascular

67

Table 27: ED CVS diagnoses – GP cohort (n=112)

Diagnosis n Specific diagnosis n ACS 20 AMI 9

Unstable angina 11 CHD 1 Angina NOS 1 Atherosclerosis 7 TIA 2

Hypertensive heart disease 1 Intracerebral haemorrhage 1 AAA rupture 2 Acute vascular disorder of intestine 1

Arrhythmia 9 SVT 2 Atrial fibrillation 3 AV block 2nd degree 1 Bradycardia 2 Cardiac arrhythmia - other 1

Other CVS diagnosis

75 Syncope 8 Mechanical complication of cardiac device 1

Other breathing abnormality 5 Dizziness 3 Palpitations 3 Chest pain - anterior chest wall 10 Chest pain on breathing 2 Chest pain unspecified 43

None (later Dx STEMI) 1 Delirium (later Dx AMI) 1

ACS = acute coronary syndrome; AMI= acute myocardial infarction; CHD = coronary heart disease; NOS = not otherwise specified; TIA = transient ischaemic attack; AAA = abdominal aortic aneurysm; SVT = supraventricular tachycardia; AV = atrioventricular; CVS = cardiovascular; STEMI = ST elevation myocardial infarction

3.2 Admissions and procedures

There were 114 patient admissions with either an ED or hospital principal CVS

diagnosis, 40 of which were for ACS (Tables 28 and 29). This represents 66 unique

patients, or 18.6% of the cohort. While some patients had hospital admission as their

ED departure destination, this was not always reflected in hospital morbidity

information; 29 patient admissions for which the ED discharge destination was

admission had no admission data in morbidity records.

Twenty-seven admissions with cardiac diagnoses occurred within one month of the test

being performed, 10 of which had a discharge diagnosis of ACS. Ten patients had a

previous admission with a cardiac diagnosis before the date of the test, including six

with ACS, three with IHD and one with heart failure. Of the 40 ACS patient admissions,

36 had admission data available. Six patients were diagnosed with STEMI, 18 with

NSTEMI, six with AMI not otherwise specified and six with UA.

Procedures performed on admitted patients are described in Table 30. Twenty one

patients with ACS underwent angiography at that admission, with 14 undergoing

revascularisation. A total of 19 patients required intensive care unit admission. There

68

were three patients who were admitted with ACS within one month of the test who did

not undergo revascularisation, all aged 85 and above.

Table 28: Admissions with CVS diagnoses – GP cohort (n=114)

Diagnosis n (%) Number of admissions ACS 40 (35.1)

Non-ACS 74 (64.9) All CVS diagnoses 114 (100)

Admission with ED presentation ACS* 25 (21.9) Non-ACS 30 (26.3) All CVS diagnoses 55 (48.2)

Admission directly to hospital ACS 15 (13.2) Non-ACS 44 (38.6) All CVS diagnoses 59 (51.8)

Admissions within 1 month of test ACS 10 (8.8) Non-ACS 17 (14.9) All CVS diagnoses 27(23.7)

Age of patients admitted (median, IQR) ACS 66 (54,76) Non-ACS 71 (47,83) All CVS diagnoses 69 (62,83)

*5 patients with non-ACS ED diagnoses were subsequently diagnosed with ACS during hospital admission; ACS = acute coronary syndrome; CVS = cardiovascular

Table 29: Details of admission CVS diagnoses– GP cohort (n=114)

Diagnosis Detailed diagnosis n ACS STEMI 6

NSTEMI 18 AMI 6 UA* 6

CHD Angina NOS 1 CHD NOS 19 Atherosclerosis of autologous bypass graft 2

Cardiomyopathy 1 Cardiogenic shock 1 Arrhythmia VT 2

AF 3 Valvular disease Symptomatic aortic stenosis 1 Heart failure 6 Chest pain NOS 6 Admission data NR 36

*4 patients with UA had no admission data in hospital morbidity data collection; ACS = acute coronary syndrome; STEMI = ST elevation myocardial infarction; NSTEMI = Non-ST elevation myocardial infarction; AMI= acute myocardial infarction; UA = unstable angina; CHD = coronary heart disease; NOS = not otherwise specified; VT= ventricular tachycardia; AF = atrial fibrillation; NR = not recorded

Table 30: Procedures performed on admitted patients - GP cohort (n=67)

ACS CHD, no ACS No ACS or CHD Total Revascularisation 14 9 1 24 Revascularisation within 1 month of test 8 0 1 9 Angiography alone 7 13 9 29 Other CVS procedure* 4 ICU admission 19

*2 electrophysiological studies, 1 testing of defibrillator, 1 cardioversion; ACS = acute coronary syndrome; CHD = coronary heart disease; CVS = cardiovascular; ICU = intensive care unit

69

3.3 Time to first hospital presentation

In total, 92 patients (25.9%) presented to hospital at least once during the follow up

period with a CVS symptom or diagnosis, whether by attendance at ED or by direct

admission. The median time to presentation was 33 days (range, zero to 551). Twenty

patients (5.6%) presented to hospital within 48 hours of testing.

Twenty seven patients (7.6%) had at least one ACS during the follow up period, ten of

whom had the ACS within 30 days of testing. The median time to presentation was 42

days (range, zero to 498). For the six patients who presented to hospital with ACS

within 48 hours of testing, the median time from specimen collection to hospital

presentation was 382 minutes (range, 80 to 1312).

70

3.4 Adverse events

Thirteen patients (3.7%) had an ACS or other adverse event within 30 days of testing,

as summarized in Table 31. The events included one death from a CVS cause, which

occurred outside of hospital within one week of the test being performed, in a patient

aged 59 years. The troponin result on this patient was negative, and survey or other

linked data were available on this patient. There was one cardiac arrest in a patient

with CHD, one episode of cardiogenic shock and one episode of an unspecified cardiac

arrhythmia. Three ACS patients did not undergo revascularisation.

The ED cohort had follow up information at 30 days for 1575 participants. 317 patients

(20.1%) had an ACS or another adverse event within 30 days of cTn testing,

significantly more than the ten patients (3.5%) with ACS in the GP cohort (P <0.001).

There were six deaths during the total follow up period, two of which had a CVS cause.

Three deaths occurred outside of hospital. All patients had comorbidities and were

aged 85 years or older.

Table 31: Adverse events within 30 days – GP and ED cohorts

GP (n=355) ED (n=1575) ACS or other adverse event within 1 month 13 317 Types of adverse events:

Death from CVS cause 1 8 Death from uncertain cause 0 0 Cardiac arrest 1 0 Cardiogenic shock 1 0 AMI - revascularisation STEMI 1 128 NSTEMI, AMI NOS 3 AMI - no revascularisation STEMI 0 84 NSTEMI, AMI NOS 3 UA - revascularisation 3 86 UA - no revascularisation 0 Emergency revascularisation NR NR Urgent revascularisation NR NR Ventricular arrhythmia 0 NR High degree AV block 0 NR Heart failure requiring intervention* NR 11

*3 admissions with heart failure, but intervention NR. ACS = acute coronary syndrome; CVS = cardiovascular, AMI= acute myocardial infarction, STEMI = ST elevation myocardial infarction, NSTEMI = Non-ST elevation myocardial infarction, NOS = not otherwise specified; AV block = atrioventricular block

71

3.4 Outcomes for survey patients in GP cohort

Of the 124 patients with both survey and linked data, there were 45 admissions or

presentations to ED, including 18 ACS occurring in eleven patients.

Six ACS occurred within one month of the test, including one STEMI. All had pain

onset more than 48 hours prior to testing and at least one CHD risk factor.

Two of these six ACS patients in this group had positive cTn results. One had typical

pain with onset more than 48 hours before the test, three risk factors and an admission

diagnosis of NSTEMI. One had atypical pain symptoms for more than 48 hours before

the test was performed with one risk factor, and a diagnosis of STEMI. Both underwent

revascularisation at that admission.

72

CHAPTER FIVE: DISCUSSION

1. KEY FINDINGS

1.1 Clinical characteristics of GP cohort

This study’s first aim was to describe clinical characteristics of patients undergoing GP-

initiated cTn testing, hypothesising that such patients are clinically different and at a

lower risk of ACS to those who present directly to ED. The GP cohort were not all at

low risk of ACS, with a majority having at least one CHD risk factor, and one in five

having a risk factor profile placing them at high risk of ACS and associated adverse

outcomes. Furthermore, most patients had pain typical of cardiac ischaemia, and most

tests were performed within the 48 hour window where the risk of complications of ACS

is highest.

However, in comparison to the ED cohort, the GP cohort did have a significantly lower

prevalence of CHD risk factors, with the exception of diabetes, where the prevalence

was similar.

1.2 GPs’ knowledge of cTn’s use and limitations

The second aim of this study was to examine GPs’ knowledge of the use of cTn testing

and its limitations. There was insufficient information from which to make firm

conclusions about GP’s knowledge of the test’s sensitivity and specificity. Seventeen

per cent of patients had impaired kidney function, one cause of a falsely positive result.

Additionally, outcome data showed that cTn testing was performed on patients who

had previously been admitted to hospital with heart failure, another cause of a false

positive result. Unfortunately, it was not possible to assess whether these diagnoses

were known to the requesting GPs, nor whether the renal impairment was a new

finding on a test performed concurrently with the cTn test.

Regarding cTn’s sensitivity, one in five cTn tests were ordered on patients who

presented within 12 hours of symptoms when the test may be insufficiently sensitive.

No serial testing was performed.

1.3 Effect of cTn on GP estimation of ACS likelihood and management

This study found that one third of GPs felt there was an intermediate or high pre-test

likelihood of ACS as the cause of their patient’s symptoms. A significant number of

73

GPs changed their assessment of likelihood of ACS in response to the test result.

Despite this, there was no significant change in the GPs’ intended management in

response to the test result.

1.4 Outcomes of patients who underwent cTn testing in primary care

This study’s final objective was to describe outcomes of patients who had a cTn test in

general practice, specifically seeking outcomes of the occurrence of complications of

an ACS and delay in diagnosis of ACS. Adverse events occurred within one month of

testing in one in 25 patients, significantly less than in the ED cohort. An important

finding was that one in four patients with negative test results presented to hospital with

a cardiovascular symptom or diagnosis during follow up, and one in 25 presented

within 48 hours of the test. One in ten patients with negative test results were admitted

to hospital with ACS, one third of which were within one month of testing.

There were too few positive results to assess the risk of adverse or beneficial

outcomes in patients who had a positive cTn test result, as only five tests were positive,

of which only two had survey data available. Only one patient had a beneficial

outcome of a positive test result contributing to ACS diagnosis, with an increase in

estimated likelihood of ACS as a result of the test.

Regarding delay, the median time from patient presentation to specimen collection was

over one hour and the median time from specimen collection to result availability was

over two hours, although not all patients had this information supplied. In patients

presenting with ACS within 48 hours of their test, the median delay from specimen

collection to hospital presentation was over six hours.

74

2. INTERPRETATION

2.1 Clinical characteristics of GP cohort

The finding that the majority of patients undergoing cTn testing have typical pain of

ACS or pain of short duration was unexpected, and there may have been other factors

not detected by the survey that reduced the risk status of patients. Examples of such

information include ECG findings or previous negative invasive investigations for CHD.

The cumulative number of CHD risk factors in the ED cohort was significantly greater

than the GP cohort. This may be explained by high risk patients heeding earlier advice

to present to ED in the event of ACS symptoms. Certainly, there are substantial

National Heart Foundation awareness campaigns marketed to high risk patients and

their GPs about the warning signs of ACS, advising them to call emergency services in

the event of the development of ACS symptoms168. Additionally, patients with multiple

CHD risk factors are more likely to present with STEMI169,170, and patients with STEMI

or symptoms of haemodynamic instability are more likely to call emergency services

rather than present to a GP157,171. Regardless, it is concerning that cTn was requested

on GP cohort patients who were automatically at high risk of ACS and adverse

outcomes based on a past history of CHD, as prior CHD is a well-known predictor of

short term mortality due to ACS172,173.

Some tests were clearly ordered on low risk patients in response to patient request, as

has been reported elsewhere174,175. Requesting GPs commented that “the test was

mainly arranged to satisfy the patient that this was unlikely cardiac in origin”, that “the

likely diagnosis was anxiety and panic but the patient felt that the chest pain was a

prominent symptom”, and that “in this case the cTn test was expected to be negative

and it provided some reassurance to the patient”. It is worth noting that a negative test

in this context may not resolve the patient’s worry about illness, as doctors are known

to overestimate the value of testing in reassuring the patient when the probability of

serious disease is low176.

2.2 GPs’ knowledge of cTn’s use and limitations

There were instances suggesting incomplete understanding of cTn’s reduced

sensitivity early after symptom onset, shown by the lack of serial tests ordered in

patients presenting within 12 hours. This correlates with findings of other

authors149,154,167. Similarly, there were instances of the test being performed in the

presence of comorbidities that reduce the test’s specificity. Laboratory advice was not

provided to the requesting doctor with the test result about these limitations. There is

75

an opportunity for education to GPs by laboratory-initiated advice, as such notifications

have been shown consistently to alter GP practice177.

2.3 Effect of cTn on GP estimation of ACS likelihood and management

cTn testing occurred most often in the setting of a less than 5% pre-test likelihood of

ACS, to further reduce the probability of ACS as a diagnosis. This is consistent with

other literature stating that cTn testing in primary care substantially reduces the clinical

likelihood of ACS178, with the findings of this this study having the added weight of

statistical significance. Interestingly, when a GP retained an intermediate or high post-

test likelihood in the face of a negative test result, the patient did not have a high risk

clinical picture; few patients in this setting had typical pain, multiple risk factors or the

recent onset of symptoms. It is possible that there were other patient factors not

detected by the survey that were influencing the GP to maintain a higher estimation of

likelihood in these situations, such as a low pre-test probability of other, non-ACS

diagnoses.

In contrast to studies in the systematic review156,179, this study showed that cTn test

results did not significantly affect GP management. Perhaps requesting GPs may have

been practising in a defensive manner, out of a fear of litigation. A cTn test is likely to

be ordered in this context given that a high proportion of medical negligence claims

relate to myocardial infarction180, and that the fear of litigation is a powerful motivator

for diagnostic test requesting177,181.

2.4 Outcomes of patients who underwent cTn testing in primary care

Over the course of the study, rates of ACS and other adverse outcomes were higher

than reported elsewhere150, though the period of follow up was longer. Although cTn

tests were ordered on patients having some risk of short term adverse outcomes, such

outcomes did not eventuate in large numbers in the period of this study, with few

positive tests or adverse outcomes within one month of the test. There were too few

deaths for a meaningful comparison of mortality rates in this study with other reports in

the literature.

The rate of presentation to hospital with CVS symptoms was high, and a number of

patients presented within hours of a negative test, suggesting the negative test result

may not have been reassuring for the patient or the GP. This finding also underscores

the difficulty in excluding serious causes of CVS symptoms in the community – even if

76

a negative cTn is taken to exclude ACS, other serious causes of CVS symptoms

requiring hospital assessment remain in the differential diagnosis114,182-184.

The admission rate for patients with negative test results was higher than expected

from the literature, with admissions in 19% of the patients in this study, compared with

14 to 16% elsewhere150,151. Possible reasons for the higher admission rate in this study

include a longer period of follow up, and a higher number of patients who were

admitted with unspecified chest pain and subsequently discharged. It was not clear if

the admissions in this study were to short stay emergency observation wards. These

wards are used increasingly since the 2009 introduction of the Four Hour Rule in Perth

tertiary hospital EDs, where a target of 85% of patients presenting to ED would be

either discharged home or admitted to a ward within four hours of presentation185.

Previously, some patients would have remained in ED for their period of observation,

and therefore not been classed as admitted. The studies reporting lower rates were

located in cities where the Four Hour Rule had not yet been implemented.

There was evidence of delay at multiple points for patients undergoing cTn testing. This

is despite guidance from laboratories to GPs about the need to avoid delay in cTn

testing, and despite recommendations from pathology bodies about the need for timely

communication of cTn test results to the requesting doctor120,186. Firstly, the median

time from patient presentation to specimen collection was over one hour, where the

interval from patient presentation to specimen collection was supplied. Even allowing

for time taken to complete the consultation and to undergo an ECG (if one was

performed) this suggests an under-appreciation of the urgent nature of the test by

either the patient or the GP, and possibly no triage for test urgency performed at the

specimen collection centre. Notably, laboratory specimen collection guides, where

available, do not indicate the urgent nature of cTn testing187.

Secondly, patients tended to present to the GP in the early part of the day, but wait

some time before having blood collected, as most specimens were collected after 12

midday. In these cases it is likely that results would not be available until outside

surgery hours. This poses a risk to the patient if systems are not in place for handling

such results, and indeed deaths have occurred in the absence of such systems118.

Thirdly, there was potential for pre-hospital delay due to laboratory factors, with the

median time from specimen collection to result availability being over two hours. This is

substantially outside the recommended 60 minute period suggested by National Heart

77

Foundation guidelines27. For the most part, though, turnaround times were less than

60 minutes where the laboratory was co-located with the specimen collection centre.

This suggests that much of the delay due to laboratory factors relates to the

geographical location of the specimen collection centre at which the patient chooses to

present, which is of course out of the control of the laboratory.

Finally, the median time from specimen collection to hospital presentation was over six

hours in patients presenting with ACS within 48 hours of their test. This is comparable

to delay reported elsewhere153. Again, this falls outside the recommended two hour

interval from presentation to reperfusion, if reperfusion is indicated10,27,42. Admittedly,

none of the patients in this study who presented with ACS within 48 hours were

ultimately diagnosed with STEMI, making early reperfusion less critical to their long

term outcomes.

78

3. STRENGTHS AND LIMITATIONS

3.1 Strengths

A strength of this research was consecutive recruitment of an unselected group of GPs

requesting cTn tests in order to reduce the risk of selection bias, as opposed to use of

a database of selected GPs in other studies179. Although the laboratories providing

results were only two of five laboratories in Perth, posing a risk of selection bias, it was

confirmed before study commencement that collection centres were distributed evenly

over the metropolitan area. It was also noted that one laboratory (St John of God)

required a patient co-payment which may have led to participants from this laboratory

being of a higher socioeconomic status. With this in mind, a government funded

laboratory (PathWest) was used for the second laboratory.

Measurement bias was minimised in the survey design by the provision of an explicit

definition of clinical factors, such as family history, smoking status and the nature and

location of pain. Using linked data sources to obtain outcome data, rather than relying

upon patient or GP recall, also increased the precision of results.

3.2 Limitations

3.2.1 Selection bias

The survey response rate was not as high as predicted, lower than in other

studies178,179. The construction of the survey may be responsible for this, as it included

a question about events in the next 30 days. Some GPs may have been waiting for the

30 days interval to lapse before completing it, and GP survey response rates are

known to decrease over time188. All GPs volunteering to take part in this study were

followed up once in order to increase response rates. A third contact would have been

valuable, as two follow up contact attempts have been shown to be more effective at

increasing response rates than a single follow up189. Other proven methods to increase

response could have been the use of a mixed mode survey, such as a response tool

built into electronic medical record result systems189. This would have avoided relying

upon the manual work involved in responding to a facsimile. The limited resources

available for this study did not permit the additional time taken for a second follow up,

nor design of an electronic tool.

It would have been ideal to compare the GP cohort with a group of patients who initially

presented to their GP but were referred directly to ED. The ED dataset was chosen for

its comparable geographic data, demographics and outcome measures, but referral

source was not collected as part of that dataset. No local research collecting this

79

information could be identified, and budgetary constraints precluded collection of ED

data specifically for this study.

The linked data had some omissions. Admission data were not provided for some

patients who had hospital admission as the ED discharge destination. Attempts were

made to track such patients using age and date of presentation but this was not

completely effective.

Most importantly, the number of adverse outcomes in the linked data was very small.

The retrospective study recorded approximately 35 urban tests per week with 5% of

these positive. This led to an initial assessment that an appropriately sized study was

possible. It is likely that many of the results in the retrospective dataset which were

classified as urban were in fact from Bunbury, a larger regional centre. At the time of

designing the study, it was planned to include patients in Bunbury. That centre is

served by a tertiary emergency department and has access to interventional cardiology

services, so was felt to be sufficiently similar to tertiary centres in urban Perth.

However, once the prospective study was underway it soon became apparent that the

patients attending collection centres in Bunbury lived in outlying rural areas. As a

result, the GP’s decision making was likely to be influenced by the patient’s increased

travel time, and therefore not comparable to that of urban GPs. Had this issue been

recognised in the design stage of the study, additional efforts would have been made to

recruit patients from other urban pathology providers.

3.2.2 Information bias

There were issues with the accuracy of time intervals recorded as part of the

assessment of delay. No objective measurements of time intervals were obtained

beyond those recalled by the patients. Many GPs provided the same time for

presentation and specimen collection, and with hindsight this wording was not clear on

the survey. Also in respect of time measurement, the interval between result availability

and the GP taking action upon the results was not measured on the survey, and it

would have been valuable to do this in order to assess the component of pre-hospital

delay due to doctor behaviour.

The accuracy of clinical information was suboptimal, as there was no corroboration of

presenting symptoms as reported by the patient to the GP. There were also incomplete

data fields for risk factors in some of the surveys from one laboratory, which in

retrospect related to an error in the paper layout of the survey that was subsequently

80

corrected. Independent verification of questionnaire responses for clinical information

and time intervals would have reduced the risk of this source of bias, but it was decided

not to proceed with independent verification in the interests of confidentiality.

It would have been valuable to extract information on potential confounders, such as

diagnoses causing falsely positive cTn results. Review of the outcome data revealed

patients undergoing cTn testing who had previously been admitted to hospital with

heart failure, but it was not clear from the data collections whether this was chronic

heart failure, nor whether this diagnosis was known to requesting GPs. While some

patients had renal impairment demonstrated by a reduced eGFR, it was not possible to

assess whether this was a new finding at the same time as cTn testing or whether this

information was already known. Only a minority of patients had previously documented

renal function tests from that laboratory. It was decided while designing the survey that

adding an exhaustive list of causes of false positive cTn would have lengthened the

survey unacceptably. Despite this, many causes of non-coronary cTn on a

comprehensive list are of very low frequency in GP, so an edited list of conditions likely

to present in primary care could have been included.

The survey did not collect ECG results. Again, it was felt that this would lengthen the

survey unacceptably and discourage participation. Additionally, the reliability of GP

interpretation of ECG data is not certain92. The task of providing the original ECG rather

than an interpretation was felt to be too onerous for the GP and also posed a risk of

patient identification. Similarly, quantitative data on blood pressure or lipid levels was

not sought from GPs, due to the effort required to provide this information being a

possible disincentive to participate in the study. Without this information, it was not

possible to apply measures of cardiovascular risk such as the TIMI score or the

National Health Foundation risk assessment criteria and the absence of these

measures of risk limits the populations to which this study’s data can be compared.

In regard to the accuracy of the linked data, it was not possible to classify outcomes

fully according to standardised data definitions for ACS research. As an example,

standardised definitions only include heart failure as an adverse outcome if it requires

intervention. Identification of this subgroup would have required access to medication

charts in hospital records, compromising patient confidentiality. There were also

instances where CHD was not accurately classified as ACS or otherwise, instead being

classed as CHD or angina not otherwise specified. This is likely to have led to

underestimation of the true numbers of ACS, though the scale of this is small.

81

4. GENERALISABILITY AND FUTURE IMPLICATIONS

4.1 Generalisability of findings

The findings of this study have important implications for patient safety. It shows that

there are inherent delays in GP-initiated cTn testing. Some of these delays can be

shortened, for example by GPs directing patients to a collection centre located close to

a laboratory. Further, GPs could impress upon patients the need to have their sample

collected as soon as practicable, and must arrange to follow up results in a timely

manner. However, even with optimal practices, some delay is unavoidable.

This study also shows that patients undergoing GP cTn testing often have CHD risk

factors. The prevalence of obesity, diabetes and hyperlipidaemia in general practice

has risen since this study was undertaken, and the presence of these risk factors

substantially increases the likelihood of short term ACS complications. It follows, then,

that the risk of ACS adverse outcomes in the GP population may also rise in the

future190-192. Although this study was too small to identify adverse outcomes occurring

early in the course of the ACS caused by a period of pre-hospital delay, such adverse

outcomes may well occur eventually.

There is a clear need to inform GPs about the limited sensitivity of cTn testing, as

based on this study understanding may be suboptimal. A negative cTn test does not

confer freedom from ACS, as shown by the finding that one in 25 patients with a

negative cTn was admitted with ACS within a month of the test. This finding supports

results of other research which state that patients deemed at low risk of ACS still carry

a residual risk of subsequent events193. GPs have been known to underestimate five-

year risk of a cardiovascular event on clinical grounds194, and must be particularly

careful to avoid equating a negative cTn test result with the absence of CHD. Instead,

they should proceed with more sensitive investigations for CHD if there is any clinical

uncertainty, and clearly communicate safety netting strategies to the patient87.

4.2 Future use of cTn testing in primary care

Point of care testing (POCT) for cTn is an attractive option which might reduce delay

and its attendant risks. There is evidence showing the efficacy of POCT in control of

chronic conditions such as diabetes and conditions requiring anticoagulation110.

Australian GPs are certainly keen to use POCT for cTn, with 43% of GPs in a 2013

survey indicating they would make use of the test if it were available195. However, a

recent systematic review of POCT of cTn levels in the pre-hospital setting reported

82

insufficient sensitivity of the test, with a negative predictive value of 57-95%108. In

addition, there is currently no Medicare rebate for POCT for cTn in general practice,

further limiting its uptake in the short to medium term196.

Another issue that emerges from this study’s findings is that cTn is ordered frequently

on patients with conditions that cause a detectable highly sensitive cTn level (hsTn).

Some authors have foreshadowed a plague of ‘troponinitis’ from indiscriminate hsTn

use in hospital patients22,197,198. It is also known that there is a progressive increase in

rates of GP pathology test ordering in general124. Consequently, as laboratory uptake of

hsTn becomes widespread, the rates of positive hsTn tests in primary care are likely to

increase. Unnecessary hospital referrals may result, if referral is initiated by GPs in

response to a single elevated value, which seems likely based on the suboptimal rates

of serial testing in this and other studies150,167. While recent research states that hsTn

does not increase the diagnosis of ACS and the burden of possible ACS patients in

emergency departments199 , that work was conducted in a controlled hospital setting

with easy access to testing protocols and a frequent throughput of ACS patients, such

that requesting hospital doctors would be very familiar with the hsTn testing algorithm

used at that site. GPs appear to require education about the need for serial cTn testing

in order to become equally familiar with the process of hsTn testing in the community.

4.3 Future implications for primary care research

This study has shown that conducting primary care research on an adequate scale on

low prevalence conditions such as ACS is challenging. For example, the ideal design to

address the objectives of this study would be randomisation of general practices to

either having access to cTn testing or not having access to cTn testing. However, an

appropriately sized study of this design would require many thousands of patients.

Moreover, it is difficult to gather clinical data on low prevalence conditions by survey,

as GP response rates to surveys are low, often 30% or less200. Recruitment rates for

survey based studies are falling, with time and workload pressures cited as

reasons188,189. Non-monetary incentives do not appear to influence response rates189.

In future investigation of low prevalence conditions, it might be possible to increase

sample sizes by using computerised research databases. The National Prescribing

Service (NPS) MedicineInsight program201 and the UK Clinical Practice Research

Datalink202 are examples of established multicentre groups which contribute data to a

common database. However, the routine data collected do not include textual

descriptions of the nature of the clinical presentation, as was required for this study.

83

One option is to collect more detailed information from selected practices within such

networks87. Research could be conducted concurrently into the diagnostic approach to

other serious low prevalence causes of chest pain, such as pulmonary embolism and

use of D-dimer testing. Another method is the use of a web-based survey. GPs could

be invited to participate by the laboratory when they check electronic results, with any

cTn test result including a link to the survey. Advantages would be a shorter response

time, improved accuracy of data around timing of presentation, and the ease of

obtaining clinical information, as the GP would be accessing the patient’s electronic

record in order to note and action the result. There is the potential for coverage error as

not all practices use electronic medical records. This effect would be small, with at least

93% of GPs using computers for pathology test ordering based on 2012 data203.

Security of patient data would need to be considered if pursuing this approach.

84

5. CONCLUSION

This study concludes that GP use of cTn testing for ACS diagnosis in primary care

does not meet the definition of appropriateness on a number of grounds. cTn is

ordered on patients at high risk of ACS and adverse outcomes of ACS occur in patients

who undergo GP-initiated cTn testing, including some events which occur within days

of testing. cTn testing adds to delay in hospital presentation, though whether this delay

contributes to adverse outcomes remains unclear. Furthermore, GP-initiated cTn

testing has not been shown to be beneficial by detecting unexpected ACS, due to low

numbers of positive tests. Most importantly, a negative cTn test is not reassuring and

does not equate to no risk of ACS.

There are potential gaps in GP understanding of cTn testing’s limitations. Additionally,

GP management of patients with possible ACS does not change as a result of cTn

testing, despite influencing estimation of likelihood of ACS. While the test is most often

used to justify GP management rather than hospital referral, this benefit is likely to be

small, as the ED burden of general practice patients is not great in any case96.

GPs are left in a difficult situation. The consequences of missing a diagnosis of ACS

can be grave, yet there are no reliable clinical predictors of ACS81,82,84, and the

investigations available in primary care have limitations, as this study shows. At the

same time, GPs have an important role as gatekeepers of the health system204-206. A

failure to accept any uncertainty may lead to unnecessary investigation and referral,

which themselves are potential causes of patient harm and health system costs.

What a GP can do is to employ other strategies in the diagnosis of ACS, such as their

their initial impression or “gut feeling”, their knowledge of the cumulative effect of

clinical risk factors, and the use of safety netting, all of which have supporting

evidence. Additionally, they should counsel patients at risk to present to hospital in the

event of ACS symptoms, and initiate primary and secondary prevention strategies for

CHD. In this way, GPs will still fulfil their key role in ACS management without the use

of cTn. The appropriate use of cTn testing for ACS diagnosis in primary care, then,

may well be to not use cTn tests at all.

85

REFERENCES

1. Australian Institute of Health and Welfare 2014. Acute coronary syndrome: validation of the method used to monitor incidence in Australia. A working paper using linked hospitalisation and deaths data from Western Australia and New South Wales. CVD 68. Canberra; 2014. 2. Nedkoff L, Briffa T, Preen D, Sanfilippo F, Hung J, Ridout S, et al. Age- and Sex-Specific Trends in the Incidence of Hospitalized Acute Coronary Syndromes in Western Australia. Circulation-Cardiovascular Quality and Outcomes 2011 Sep;4(5):557-564. 3. Briffa T, Hickling S, Knuiman M, Hobbs M, Hung J, Sanfilippo FM, et al. Long term survival after evidence based treatment of acute myocardial infarction and revascularisation: follow-up of population based Perth MONICA cohort, 1984-2005. Br Med J 2009 Jan;338:b36 (approx. 8 p.). 4. Australian Institute of Health and Welfare. Monitoring acute coronary syndrome using national hospital data: an information paper on trends and issues. 2011;Cat. no. CVD 57. Canberra: AIHW. 5. Deloitte Access Economics. ACS in perspective: the importance of secondary prevention. Canberra: Deloitte Access Economics; 2011 [cited 2015 8 February]. Available from: https://www.deloitteaccesseconomics.com.au/uploads/File/Final%20Report%20ACS%20in%20Perspective%20Nov%202011.pdf 6. Than M, Cullen L, Reid CM, Lim SH, Aldous S, Ardagh MW, et al. A 2-h diagnostic protocol to assess patients with chest pain symptoms in the Asia-Pacific region (ASPECT): a prospective observational validation study. Lancet 2011 Mar-Apr;377(9771):1077-1084. 7. Roffi M, Patrono C, Collet JP, Mueller C, Valgimigli M, Andreotti F, et al. 2015 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation. Eur Heart J. 2016 Jan 14;37(3):267-315. 8. Aroney CN, Aylward P, Kelly AM, Chew DPB, Clune E. National Heart Foundation of Australia Cardiac Society of Australia and New Zealand Guidelines for the management of acute coronary syndromes 2006. Med J Aust 2006 Apr 17;184(8):S1-S30. 9. Anderson JL, Adams CD, Antman EA. ACC/AHA 2007 guidelines for the management of patients with unstable angina/non-ST-elevation myocardial infarction: A report of the American College of Cardiology/American Heart Association task force on practice guidelines. J Am Coll Cardiol 2008 March;51(9):974-974. 10. Van de Werf F, Bax J, Betriu A, Blomstrom-Lundqvist C, Crea F, Falk V, et al. Management of acute myocardial infarction in patients presenting with persistent ST-segment elevation. Eur Heart J 2008 Dec;29(23):2909-2945. 11. Thygesen K, Alpert JS, Jaffe AS, Simoons ML, Chaitman BR, White HD, et al. Third Universal Definition of Myocardial Infarction. Circulation 2012 Oct;126(16):2020-2035. 12. Canto JG, Fincher C, Kiefe CI, Allison JJ, Li Q, Baker S, et al. Atypical presentations among Medicare beneficiaries with unstable angina: Is it time to redefine the classical clinical presentation? J Am Coll Cardiol 2001 Feb;37(2):376A-376A. 13. Pope JH, Aufderheide TP, Ruthazer R, Woolard RH, Feldman JA, Beshansky JR, et al. Missed diagnoses of acute cardiac ischemia in the emergency department. N Engl J Med 2000 Apr 20;342(16):1163-1170. 14. Sheifer SE, Rathore SS, Gersh BJ, Weinfurt KP, Oetgen WJ, Breall JA, et al. Time to presentation with acute myocardial infarction in the elderly - Associations with race, sex, and socioeconomic characteristics. Circulation 2000 Oct 3;102(14):1651-1656. 15. White HD, Chew DP. Acute myocardial infarction. Lancet 2008 Aug 16;372(9638):570-584.

http://www.deloitteaccesseconomics.com.au/uploads/File/Final%20Report%20ACS%20in%20Perspective%20Nov%202011.pdf

http://www.deloitteaccesseconomics.com.au/uploads/File/Final%20Report%20ACS%20in%20Perspective%20Nov%202011.pdf

86

16. Walters DL, Cunningham C. Managing acute coronary syndromes in the prehospital and emergency setting: New guidelines from the Australian Resuscitation Council and New Zealand Resuscitation Council. Emerg Med Australas 2011 Jun;23(3):240-243. 17. Body R, Carley S, Wibberley C, McDowell G, Ferguson J, Mackway-Jones K. The value of symptoms and signs in the emergent diagnosis of acute coronary syndromes. Resuscitation 2010 Mar;81(3):281-286. 18. Loten C, Isbister G, Jamcotchian M, Hullick C, MacElduff P, Attia J, et al. Adverse outcomes following emergency department discharge of patients with possible acute coronary syndrome. Emerg Med Australas 2009 Dec;21(6):455-464. 19. Han JH, Lindsell CJ, Storrow AB, Luber S, Hoekstra JW, Hollander JE, et al. The role of cardiac risk factor burden in diagnosing acute coronary syndromes in the emergency department setting. Ann Emerg Med 2007 Feb;49(2):145-152. 20. Rogers WJ, Canto JG, Lambrew CT, Tiefenbrunn AJ, Kinkaid B, Shoultz DA, et al. Temporal trends in the treatment of over 1.5 million patients with myocardial infarction in the US from 1990 through 1999 - The National Registry of Myocardial Infarction 1, 2 and 3. J Am Coll Cardiol 2000 Dec;36(7):2056-2063. 21. Antman EM, Tanasijevic MJ, Thompson B, Schactman M, McCabe CH, Cannon CP, et al. Cardiac-specific troponin I levels to predict the risk of mortality in patients with acute coronary syndromes. N Engl J Med 1996 Oct 31;335(18):1342-1349. 22. Cullen L. Troponin: A risk-defining biomarker for emergency department physicians. Emerg Med Australas 2011 Aug;23(4):391-394. 23. Falahati A, Sharkey SW, Christensen D, McCoy M, Miller EA, Murakami MA, et al. Implementation of serum cardiac troponin I as marker for detection of acute myocardial infarction. Am Heart J 1999 Feb;137(2):332-7. 24. Ghaemmaghami CA, Brady WJ. Pitfalls in the emergency department diagnosis of acute myocardial infarction. Emerg Med Clin North Am 2001 May;19(2):351-369. 25. Reece JB, Campbell NA. Campbell Biology. 9th ed: Pearson Education Australia; 2011. 26. Aldous SJ. Cardiac biomarkers in acute myocardial infarction. Int J Cardiol 2013 Apr 15;164(3):282-294. 27. Chew DP, Aroney CN, Aylward PE, Kelly AM, White HD, Tideman PA, et al. 2011 Addendum to the National Heart Foundation of Australia/Cardiac Society of Australia and New Zealand Guidelines for the Management of Acute Coronary Syndromes (ACS) 2006. Heart Lung and Circulation 2011 Aug;20(8):487-502. 28. Thygesen K, Mair J, Giannitsis E, Mueller C, Lindahl B, Blankenberg S, et al. How to use high-sensitivity cardiac troponins in acute cardiac care. Eur Heart J 2012 Sep;33(18):2252–2257. 29. Bima A, Sikaris K. Towards appreciating appropriate clinical responses to highly sensitive cardiac troponin assays. Int Med J 2012 Oct;42(10):16-22. 30. Heidenreich PA, Alloggiamento T, Melsop K, McDonald KM, Go AS, Hlatky MA. The prognostic value of troponin in patients with non-ST elevation acute coronary syndromes: A meta-analysis. J Am Coll Cardiol 2001 Aug;38(2):478-485. 31. James SK, Lindahl B, Siegbahn A, Stridsberg M, Venge P, Armstrong P, et al. N-terminal pro-brain natriuretic peptide and other risk markers for the separate prediction of mortality and subsequent myocardial infarction in patients with unstable coronary artery disease - A Global Utilization of Strategies to Open occluded arteries (GUSTO)-IV substudy. Circulation 2003 Jul 22;108(3):275-281. 32. Flindell JA, Finn JC, Gibson NP, Jacobs IG. Short-term risk of adverse outcome is significantly higher in patients returning an abnormal troponin result when tested in the emergency department. Emerg Med Australas 2009 Dec;21(6):465-471. 33. MacDonald PS, Newton PJ, Davidson PM. The SNAPSHOT ACS study: getting the big picture on acute coronary syndrome. Med J Aust 2013 Aug;199(3):147-148.

87

34. Omland T. New sensitive cardiac troponin assays for the early diagnosis of myocardial infarction. Drugs of Today 2011 Apr;47(4):303-12. 35. Kavsak PA, Wang X, Ko DT, MacRae AR, Jaffe AS. Short- and long-term risk stratification using a next-generation, high-sensitivity research cardiac troponin I (hs-cTnI) assay in an emergency department chest pain population. Clin Chem 2009 Oct;55(10):1809-15. 36. Lindahl B, Venge P, James S. The new high-sensitivity cardiac troponin T assay improves risk assessment in acute coronary syndromes. Am Heart J 2010 Aug;160(2):224-229. 37. deFilippi CR, de Lemos JA, Christenson RH, Gottdiener JS, Kop WJ, Zhan M, et al. Association of Serial Measures of Cardiac Troponin T Using a Sensitive Assay With Incident Heart Failure and Cardiovascular Mortality in Older Adults. JAMA 2010 Dec 8;304(22):2494-2502. 38. Jeremias A, Gibson CM. Narrative review: Alternative causes for elevated cardiac troponin levels when acute coronary syndromes are excluded. Ann Int Med 2005 May 3;142(9):786-791. 39. Aldous SJ. The highs and lows of high sensitivity troponin. Int Med J 2011 Jul;41(7):513-515. 40. Department of Health Western Australia. The Model of Care for Acute Coronary Syndromes in Western Australia. Perth: Health Networks Branch Department of Health Western Australia; 2009. 41. Chew DPB, Allan RM, Aroney CN, Sheerin NJ. National data elements for the clinical management of acute coronary syndromes. Med J Aust 2005;182(9 Suppl):S1-14. 42. O'Gara PT, Kushner FG, Ascheim DD, Casey DE, Chung MK, de Lemos JA, et al. 2013 ACCF/AHA Guideline for the Management of ST-Elevation Myocardial Infarction A Report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 2013 Jan;127(4):E362-E425. 43. Gersh BJ, Stone GW, White HD, Holmes DR. Pharmacological facilitation of primary percutaneous coronary intervention for acute myocardial infarction - Is the slope of the curve the shape of the future? JAMA 2005 Feb;293(8):979-986. 44. De Luca G, Suryapranata H, Ottervanger JP, Antman EM. Time delay to treatment and mortality in primary angioplasty for acute myocardial infarction: every minute of delay counts. Circulation 2004 Mar 16;109(10):1223-5. 45. Aldous SJ, Florkowski CM, Crozier IG, Elliott J, George P, Lainchbury JG, et al. Comparison of high sensitivity and contemporary troponin assays for the early detection of acute myocardial infarction in the emergency department. Ann Clin Biochem 2011 May;48:241-8. 46. Amsterdam EA, Kirk JD, Bluemke DA, Diercks D, Farkouh ME, Garvey JL, et al. Testing of Low-Risk Patients Presenting to the Emergency Department With Chest Pain A Scientific Statement From the American Heart Association. Circulation 2010 Oct;122(17):1756-1776. 47. Hawkins RC. Laboratory turnaround time. The Clinical Biochemist Reviews - Australian Association of Clinical Biochemists 2007;28(4). 48. Wu AHB, Apple FS, Gibler WB, Jesse RL, Warshaw MM, Valdes R. National Academy of Clinical Biochemistry standards of laboratory practice: Recommendations for the use of cardiac markers in coronary artery diseases. Clin Chem 1999 Jul;45(7):1104-1121. 49. Than M, Cullen L, Aldous S, Parsonage WA, Reid CM, Greenslade J, et al. 2-Hour Accelerated Diagnostic Protocol to Assess Patients With Chest Pain Symptoms Using Contemporary Troponins as the Only Biomarker - The ADAPT Trial. J Am Coll Cardiol 2012 Jun 5;59(23):2091-2098.

88

50. Goodacre SW, Bradburn M, Cross E, Collinson P, Gray A, Hall AS, et al. The Randomised Assessment of Treatment using Panel Assay of Cardiac Markers (RATPAC) trial: a randomised controlled trial of point-of-care cardiac markers in the emergency department. Heart 2011 Feb;97(3):190-196. 51. Aldous SJ, Richards M, Cullen L, Troughton R, Than M. A 2-hour thrombolysis in myocardial infarction score outperforms other risk stratification tools in patients presenting with possible acute coronary syndromes: comparison of chest pain risk stratification tools. Am Heart J;164(4):516-23. 52. Palamalai V, Murakami MM, Apple FS. Diagnostic performance of four point of care cardiac troponin I assays to rule in and rule out acute myocardial infarction. Clin Biochem 2013 Nov;46(16-17):1631-1635. 53. Aldous SJ. Cardiac biomarkers in acute myocardial infarction. Int J Cardiol 2013;164(3):282-94. 54. Drew BJ, Califf RM, Funk M, Kaufman ES, Krucoff MW, Laks MM, et al. Practice standards for electrocardiographic monitoring in hospital settings - An American Heart Association Scientific Statement from the councils on cardiovascular nursing, clinical cardiology, and cardiovascular disease in the young. Circulation 2004 Oct;110(17):2721-2746. 55. Akkerhuis KM, Klootwijk PAJ, Lindeboom W, Umans V, Meij S, Kint PP, et al. Recurrent ischaemia during continuous multilead ST-segment monitoring identifies patients with acute coronary syndromes at high risk of adverse cardiac events - Meta-analysis of three studies involving 995 patients. Eur Heart J 2001 Nov;22(21):1997-2006. 56. Reichlin T, Irfan A, Twerenbold R, Reiter M, Hochholzer W, Burkhalter H, et al. Utility of absolute and relative changes in cardiac troponin concentrations in the early diagnosis of acute myocardial infarction. Circulation 2011;124(2):136-45. 57. Keller T, Zeller T, Peetz D, Tzikas S, Roth A, Czyz E, et al. Sensitive Troponin I Assay in Early Diagnosis of Acute Myocardial Infarction. N Engl J Med 2009 Aug 27;361(9):868-877. 58. Mills NL, Churchhouse AMD, Lee KK, Anand A, Gamble D, Shah ASV, et al. Implementation of a Sensitive Troponin I Assay and Risk of Recurrent Myocardial Infarction and Death in Patients With Suspected Acute Coronary Syndrome. JAMA 2011 Mar;305(12):1210-1216. 59. Reichlin T, Schindler C, Drexler B, Twerenbold R, Reiter M, Zellweger C, et al. One-Hour Rule-out and Rule-in of Acute Myocardial Infarction Using High-Sensitivity Cardiac Troponin T. Arch Intern Med 2012 Sep;172(16):1211-1218. 60. Aldous SJ, Richards MA, Cullen L, Troughton R, Than M. A New Improved Accelerated Diagnostic Protocol Safely Identifies Low-risk Patients With Chest Pain in the Emergency Department. Acad Emerg Med 2012 May;19(5):510-516. 61. Body R, Carley S, McDowell G, Jaffe AS, France M, Cruickshank K, et al. Rapid exclusion of acute myocardial infarction in patients with undetectable troponin using a high-sensitivity assay. J Am Coll Cardiol;58(13):1332-9. 62. Thokala P, Goodacre SW, Collinson PO, Stevens JW, Mills NL, Newby DE, et al. Cost-effectiveness of presentation versus delayed troponin testing for acute myocardial infarction. Heart 2012 Oct;98(20):1498-1503. 63. Parsonage WA, Cullen L, Younger JF. The approach to patients with possible cardiac chest pain. Med J Aust 2013 Jul 8;199(1):30-34. 64. Cullen L. Australasian Association of Clinical Biochemists 2012 Webinar Series - Troponin in Acute Coronary Syndrome[Internet]. 2012 [cited 2014 1 January]. Available from: http://www.aacb.asn.au/professionaldevelopment/2012-webinars/troponin-in-acute-coronary-syndrome 65. Aakre KM, Sandberg S. Can Changes in Troponin Results Be Useful in Diagnosing Myocardial Infarction? Clin Chem 2010 Jul;56(7):1047-1049.

http://www.aacb.asn.au/professionaldevelopment/2012-webinars/troponin-in-acute-coronary-syndrome

http://www.aacb.asn.au/professionaldevelopment/2012-webinars/troponin-in-acute-coronary-syndrome

89

66. Wallace TW, Abdullah SM, Drazner MH, Das SR, Khera A, McGuire DK, et al. Prevalence and determinants of troponin T elevation in the general population. Circulation 2006 Apr 25;113(16):1958-1965. 67. Saunders JT, Nambi V, de Lemos JA, Chambless LE, Virani SS, Boerwinkle E, et al. Cardiac Troponin T Measured by a Highly Sensitive Assay Predicts Coronary Heart Disease, Heart Failure, and Mortality in the Atherosclerosis Risk in Communities Study. Circulation 2011 Apr 5;123(13):1367-76. 68. Otsuka T, Kawada T, Ibuki C, Seino Y. Association between high-sensitivity cardiac troponin T levels and the predicted cardiovascular risk in middle-aged men without overt cardiovascular disease. Am Heart J 2010 Jun;159(6):972-978. 69. Apple FS. A New Season for Cardiac Troponin Assays: It's Time to Keep a Scorecard. Clin Chem 2009 Jul;55(7):1303-1306. 70. Jaffe AS, Apple FS. High-Sensitivity Cardiac Troponin: Hype, Help, and Reality. Clin Chem 2010 Mar;56(3):342-344. 71. Vasikaran SD, Macdonald SPJ, Sikaris KA. High-sensitivity cardiac troponin assays for risk stratification and for the diagnosis of acute myocardial infarction. Ann Clin Biochem 2012 May;49(3):209-210. 72. Christ M, Bertsch T. High-sensitivity troponin assays and clinical decisions. Ann Clin Biochem 2012 May;49(3):306-307. 73. Aldous SJ. Response to 'High-sensitivity troponin assays and clinical decisions' by Christ and Bertsch. Ann Clin Biochem 2012 May;49(3):307-307. 74. Twerenbold R, Jaffe A, Reichlin T, Reiter M, Mueller C. High-sensitive troponin T measurements: what do we gain and what are the challenges? Eur Heart J 2012 Mar;33(5):579-586. 75. Twerenbold R, Reichlin T, Mueller C. Clinical application of sensitive cardiac troponin assays: potential and limitations. Biomark Med 2010 Jun;4(3):395-401. 76. Meune C, Balmelli C, Twerenbold R, Reichlin T, Reiter M, Haaf P, et al. Patients with Acute Coronary Syndrome and Normal High-sensitivity Troponin. Am J Med 2011 Dec;124(12):1151-1157. 77. Reichlin T, Twerenbold R, Mueller C. Comment on "High-Sensitivity Cardiac Troponin: Hype, Help, and Reality". Clin Chem 2010 Jul;56(7):1198-1199. 78. Lozzi L, Carstensen S, Rasmussen H, Nelson G. Why do acute myocardial infarction patients not call an ambulance? An interview with patients presenting to hospital with acute myocardial infarction symptoms. Intern Med J 2005 Nov;35(11):668-671. 79. Britt H MG, Henderson J, Bayram C, Valenti L, Harrison C, et al. General practice activity in Australia 2012–13. General practice series no.33.[Internet]. Sydney: Sydney University Press; 2013 [cited 2014 9 January]. Available from: http://hdl.handle.net/2123/9365 80. Australian Institute of Health and Welfare Australian GP Statistics and Classification Centre. SAND abstract No. 188 from the BEACH program: Acute coronary syndrome among general practice patients.[Internet]. Sydney: FMRC University of Sydney; 2012 [cited 2014 December 22]. Available from: http://sydney.edu.au/medicine/fmrc/publications/sand-abstracts/index.php 81. Bruyninckx R, Aertgeerts B, Bruyninckx P, Buntinx F. Signs and symptoms in diagnosing acute myocardial, infarction and acute coronary syndrome: a diagnostic meta-analysis. Br J Gen Prac 2008 Feb;58(547):105-111. 82. Bossaert L, O'Connor RE, Arntz H-R, Brooks SC, Diercks D, Feitosa-Filho G, et al. Part 9: Acute coronary syndromes 2010 International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science with Treatment Recommendations. Resuscitation 2010 Oct;81(1).

http://hdl.handle.net/2123/9365

http://sydney.edu.au/medicine/fmrc/publications/sand-abstracts/index.php

90

83. Mant J, McManus RJ, Oakes RAL, Delaney BC, Barton PM, Deeks JJ, et al. Systematic review and modelling of the investigation of acute and chronic chest pain presenting in primary care. Health Technol Assess 2004 Feb;8(2):(approx. 174 p.). 84. Swap CJ, Nagurney JT. Value and limitations of chest pain history in the evaluation of patients with suspected acute coronary syndromes. JAMA 2005 Nov;294(20):2623-2629. 85. Antman EM, Cohen M, Bernink P, McCabe CH, Horacek T, Papuchis G, et al. The TIMI risk score for unstable angina/non-ST elevation MI - A method for prognostication and therapeutic decision making. JAMA 2000 Aug;284(7):835-842. 86. Brieger D, Fox KAA, FitzGerald G, Eagle KA, Budaj A, Avezum A, et al. Predicting freedom from clinical events in non-ST-elevation acute coronary syndromes: the Global Registry of Acute Coronary Events. Heart 2009 Jun;95(11):888-894. 87. Buntinx F, Mant D, Van den Bruel A, Donner-Banzhof N, Dinant G-J. Dealing with low-incidence serious diseases in general practice. Br J Gen Pract 2011 Jan;61(582):43-46. 88. O'Riordan M, Dahinden A, Akturk Z, Ortiz JMB, Dagdeviren N, Elwyn G, et al. Dealing with uncertainty in general practice: an essential skill for the general practitioner. Qual Prim Care 2011 2011;19(3):175-81. 89. Bosner S, Haasenritter J, Becker A, Karatolios K, Vaucher P, Gencer B, et al. Ruling out coronary artery disease in primary care: development and validation of a simple prediction rule. CMAJ 2010 Sep;182(12):1295-1300. 90. Haasenritter J, Bosner S, Vaucher P, Herzig L, Heinzel-Gutenbrunner M, Baum E, et al. Ruling out coronary heart disease in primary care: external validation of a clinical prediction rule. Br J Gen Prac 2012 Jun;62(599):e415-e421. 91. Gencer B, Vaucher P, Herzig L, Verdon F, Ruffieux C, Boesner S, et al. Ruling out coronary heart disease in primary care patients with chest pain: a clinical prediction. BMC Med 2010 Jan 21;8(9):(approx. 9 pages). 92. Whitman M, Layt D, Yelland M. Key findings on ECGs Level of agreement between GPs and cardiologists. Aust Fam Phys 2012 Jan-Feb;41(1-2):59-62. 93. Rutten FH, Kessels AGH, Willems FF, Hoes AW. Electrocardiography in primary care; is it useful? Int J Cardiol 2000 Jul 31;74(2-3):199-205. 94. Patel MR, Spertus JA, Brindis RG, Hendel RC, Douglas PS, Peterson ED, et al. ACCF proposed method for evaluating the appropriateness of cardiovascular imaging. J Am Coll Cardiol 2005 Oct 18;46(8):1606-1613. 95. Marshall GA, Wijeratne NG, Thomas D. Should general practitioners order troponin tests? Med J Aust 2014;201:155-157. 96. Sprivulis P. Estimation of the general practice workload of a metropolitan teaching hospital emergency department. Emerg Med 2003;15(1):32-7. 97. Hedges JR, Mann NC, Meischke H, Robbins M, Goldberg R, Zapka J, et al. Assessment of chest pain onset and out-of-hospital delay using standardized interview questions: The REACT pilot study. Acad Emerg Med 1998 Aug;5(8). 98. Taylor DM, Garewal D, Carter M, Bailey M, Aggarwal A. Factors that impact upon the time to hospital presentation following the onset of chest pain. Emerg Med Australas 2005 Jun;17(3):204-11. 99. Bleeker JK, Simoons ML, Erdman RA, Leenders CM, Kruyssen HA, Lamers LM, et al. Patient and doctor delay in acute myocardial infarction: a study in Rotterdam, The Netherlands. Br J Gen Prac 1995 Apr;45(393):181-4. 100. Finn JC, Bett JHN, Shilton TR, Cunningham C, Thompson PL. Patient delay in responding to symptoms of possible heart attack: can we reduce time to care? Med J Aust 2007 Sep 3;187(5):293-298. 101. Kostopoulou O, Delaney BC, Munro CW. Diagnostic difficulty and error in primary care - a systematic review. Fam Pract 2008 Dec;25(6):400-413.

91

102. Ottesen MM, Dixen U, Torp-Pedersen C, Kober L. Prehospital behaviour of patients admitted with acute coronary syndrome or witnessed cardiac arrest. Scand Cardiovasc J 2003;37(3):141-8. 103. Alonzo AA. The effect of health care provider consultation on acute coronary syndrome care-seeking delay. Heart Lung 2007 Sep-Oct;36(5). 104. Hitchcock T, Rossouw F, McCoubrie D, Meek S. Observational study of prehospital delays in patients with chest pain. Emerg Med J 2003 May;20(3):270-3. 105. GISSI Avoidable Delay Study Group. Epidemiology of avoidable delay in the care of patients with acute myocardial infarction in Italy. A GISSI-generated study. Arch Intern Med 1995 Jul 24;155(14):1481-8. 106. Hafiz AM, Naidu SS, DeLeon J, Islam S, Alkhatib B, Lorenz M, et al. Impact of first contact on symptom onset-to-door time in patients presenting for primary percutaneous coronary intervention. Am J Emerg Med 2013 Jun;31(6):922-927. 107. Royal Australian College of General Practitioners. Standards for general practices (4th edition) [Internet]. Melbourne: Royal Australian College of General Practitioners; 2010 [updated November 26, 2013; cited 2013 December 24]. Available from: http://www.racgp.org.au/your-practice/standards/standards4thedition 108. Bruins Slot MH, van der Heijden GJ, Stelpstra SD, Hoes AW, Rutten FH. Point-of-care tests in suspected acute myocardial infarction: a systematic review. Int J Cardiol 2013;168(6):5355-62. 109. Wu AHB, Christenson RH. Analytical and assay issues for use of cardiac troponin testing for risk stratification in primary care. Clin Biochem 2013 Aug;46(12):969-978. 110. Laurence CO, Gialamas A, Bubner T, Yelland L, Willson K, Ryan P, et al. Patient satisfaction with point-of-care testing in general practice. Br J Gen Pract 2010;60(572):e98-104. 111. Laurence C, Gialamas A, Yelland L, Bubner T, Ryan P, Willson K, et al. A pragmatic cluster randomised controlled trial to evaluate the safety, clinical effectiveness, cost effectiveness and satisfaction with point of care testing in a general practice setting - rationale, design and baseline characteristics. Trials 2008 Aug 6;9:[approx. 14 pages]. 112. Bosner S, Becker A, Haasenritter J, Abu Hani M, Keller H, Sonnichsen AC, et al. Chest pain in primary care: Epidemiology and pre-work-up probabilities. European Journal of General Practice 2009 Nov;15(3):141-146. 113. Erhardt L, Herlitz J, Bossaert L, Halinen M, Keltai M, Koster R, et al. Task force on the management of chest pain. Eur Heart J 2002;23(15):1153-1176. 114. Svavarsdottir AE, Jonasson MR, Gudmundsson GH, Fjeldsted K. Chest pain in family practice - Diagnosis and long-term outcome in a community setting. Can Fam Physician 1996 Jun;42. 115. Diamond GA, Kaul S. How Would the Reverend Bayes Interpret High-Sensitivity Troponin? Circulation 2010 Mar;121(10):1172-1175. 116. Fabbri A, Ottani F, Marchesini G, Galvani M, Vandelli A. Predicting unfavorable outcome in subjects with diagnosis of chest pain of undifferentiated origin. Am J Emerg Med 2012 Jan;30(1):61-67. 117. Apple FS, Ler R, Murakami MM. Determination of 19 Cardiac Troponin I and T Assay 99th Percentile Values from a Common Presumably Healthy Population. Clin Chem 2012 Nov;58(11):1574-1581. 118. Bird S. Follow up of test results after hours. Aust Fam Physician 2007 Sep;36(9):761-762. 119. Hadlow N. Troponin - DON'T wait for a result.[Internet]. 2007 [cited 2014 January 9]. Available from: www.wdp.com.au 120. Edwards G. Use of Troponin I in General Practice. SJOG Health Care Lab Update September 2008[Internet ]. 2008 [cited 2014 January 9]. Available from: www.sjog.org.au

http://www.racgp.org.au/your-practice/standards/standards4thedition

http://www.wdp.com.au/

http://www.sjog.org.au/

92

121. Clinipath Pathology. Clinipath Pathology Quality Improvement February 2011. New Assay: High Sensitivity (hs) Troponin T[Internet]. 2011 [cited 2014 January 9]. Available from: http://www.clinipathpathology.com.au/ 122. Department of Human Services (AU). Medicare Benefits Schedule [Internet]. January 2014 [cited 2014 8 January]. Available from: http://www.health.gov.au/internet/mbsonline/publishing.nsf/Content/Medicare-Benefits-Schedule-MBS-1 123. Department of Human Services (AU). Medicare Statistics[Internet]. 2014 [cited 2014 8 January]. Available from: https://www.medicareaustralia.gov.au/statistics/mbs_item.shtml 124. Bayram C, Britt H, Miller G, Valenti L. Evidence-practice gap in GP pathology test ordering: a comparison of BEACH (Bettering the Evaluation And Care of Health) pathology data and recommended testing: Final report to the Quality Use of Pathology Program. Sydney, Family Medicine Research Centre, School of Public Health,The University of Sydney; 2009. Available from: www.health.gov.au/internet/main/publishing.nsf/Content/9C300FE48F876E95CA257BF0001ACE0E/$File/Evidence-practice%20gap%20in%20GP%20pathology%20test%20ordering.pdf 125. Dynamed Editorial Team. [Internet]. Ipswich (MA) EBSCO Publishing; [cited 24 December 2013]. Available from: Dynamed 126. Xu B, MacIsaac AI. What does an elevated troponin mean?--An update on the definition of myocardial infarction. Aust Fam Phys 2013 Aug;42(8):554-9. 127. Royal College of Pathologists of Australasia. RCPA Manual [Internet]. Surry Hills: RCPA [updated 22 August 2011; cited 2014 8 January]. Available from: http://rcpamanual.edu.au/ 128. Scottish Intercollegiate Guidelines Network (SIGN). Acute coronary syndromes (SIGN publication no. 93). Edinburgh: SIGN [updated February 2013; cited 2014 January 8]. Available from: http://www.sign.ac.uk 129. National Institute for Health and Clinical Excellence. Chest pain of recent onset: Assessment and diagnosis of recent onset chest pain or discomfort of suspected cardiac origin. Manchester: NICE; 2013 [updated 19 December 2013; cited 2014 January 1]. Available from: www.nice.org.uk 130. Newby LK, Jesse RL, Babb JD, Christenson RH, De Fer TM, Diamond GA, et al. ACCF 2012 Expert Consensus Document on Practical Clinical Considerations in the Interpretation of Troponin Elevations - A Report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents. J Am Coll Cardiol 2012 Dec 11;60(23):2427-2463. 131. Smellie WSA, Forth J, Smart SRS, Galloway MJ, Irving W, Bareford D, et al. Best practice in primary care pathology: review 7. J Clin Pathol 2007 May 1;60(5):458-465. 132. National Health and Medical Research Council. How to review the evidence: systematic identification and review of the scientific literature. Canberra: Commonwealth of Australia; 2000. Available from: http://www.nhmrc.gov.au/guidelines/publications/cp65 133. Centre for Reviews and Dissemination. Systematic Reviews: CRD’s guidance for undertaking reviews in health care. York: York Publishing Services; 2011. Available from: http://www.york.ac.uk/inst/crd/index_guidance.htm 134. Popay J RH, Sowden A, Petticrew M, Arai L, Rodgers M, Britten N, Roen K, Duffy S. Guidance on the conduct of narrative synthesis in systematic reviews. A product from the ESRC methods programme[Internet]. Lancaster: Institute of Health Research: ESRC Methods Program; 2006 [cited 2014 27 January]. Available from: http://www.researchgate.net/ 135. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. Ann Intern Med 2009 Aug 18;151(4):264-W64.

http://www.clinipathpathology.com.au/

http://www.health.gov.au/internet/mbsonline/publishing.nsf/Content/Medicare-Benefits-Schedule-MBS-1

http://www.health.gov.au/internet/mbsonline/publishing.nsf/Content/Medicare-Benefits-Schedule-MBS-1

http://www.medicareaustralia.gov.au/statistics/mbs_item.shtml

http://www.health.gov.au/internet/main/publishing.nsf/Content/9C300FE48F876E95CA257BF0001ACE0E/$File/Evidence-practice%20gap%20in%20GP%20pathology%20test%20ordering.pdf



http://rcpamanual.edu.au/

http://www.sign.ac.uk/

http://www.nice.org.uk/

http://www.nhmrc.gov.au/guidelines/publications/cp65

http://www.york.ac.uk/inst/crd/index_guidance.htm

http://www.researchgate.net/

93

136. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JPA, et al. The PRISMA Statement for Reporting Systematic Reviews and Meta-Analyses of Studies That Evaluate Health Care Interventions: Explanation and Elaboration. Ann Intern Med 2009 Aug 18;151(4):W65-W94. 137. Cullen L, Than M, Brown AFT, Richards M, Parsonage W, Flaws D, et al. Comprehensive standardized data definitions for acute coronary syndrome research in emergency departments in Australasia. Emerg Med Australas 2010 Feb;22(1):35-55. 138. Bruyninckx R, Van den Bruel A, Buntinx F, Van Casteren V, Aertgeerts B. Excess of mortality in patients with chest pain peaks in the first 3 days period after the incident and normalizes after 1 month. Fam Pract 2010 Dec;27(6):604-8. 139. Centre for Reviews and Dissemination. In: DARE (Database of Abstracts of Reviews of Effects) [internet]. York: Centre for Reviews and Dissemination, University of York. 140. The Cochrane Library. Chichester: Wiley; Issue 12 of 12, December 2013. 141. National Institute for Health and Clinical Excellence (UK). NICE guidance [Internet]. London: National Institute for Health and Care Excellence; 2013 [cited 2013 30 December]. Available from: http://www.nice.org.uk 142. National Institute for Health Research (NIHR). NIHR Journals Library [Internet]. NIHR Evaluation, Trials and Studies Coordinating Centre; 2013 [cited 2013 30 December]. Available from: http://www.journalslibrary.nihr.ac.uk/ 143. Agency for Healthcare Research and Quality (AHRQ) (US). National Guidelines Clearinghouse[Internet]. Rockville, MD U.S. Department of Health and Human Services; 2013 [cited 2013 30 December]. Available from: http://guidelines.gov/ 144. Scottish Intercollegiate Guidelines Network (SIGN). [Internet]. Edinburgh: Scottish Intercollegiate Guidelines Network; 2013 [cited 2013 30 December]. Available from: www.sign.ac.uk/index.html 145. EndNote X6. [computer program]. Carlsbad, CA: Thomson Reuters; 2012. 146. Higgins JPT, Green S, (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0[Internet]. The Cochrane Collaboration; 2011 [updated 2011 March; cited 2015 25 January]. Available from: http://www.cochrane.org/handbook 147. CASP UK. Critical Appraisal Skills Programme: making sense of evidence about clinical effectiveness: 10 questions to help you make sense of qualitative research.[Internet]. Oxford, United Kingdom; 2013 [updated 31 May 2013; cited 2013 9 July]. Available from: http://www.casp-uk.net/wp-content/uploads/2011/11/CASP-Qualitative-Research-Checklist-31.05.13.pdf 148. Dixon-Woods M, Agarwal S, Jones D, Young B, Sutton A. Synthesising qualitative and quantitative evidence: a review of possible methods. J Health Serv Res Policy 2005;10(1):45-53. 149. Mann S, Tietjens J, Law K, Elley R. Troponin testing for chest pain in primary healthcare: a New Zealand audit. N Z Med J 2006;119(1238):U2083(approx. 8 p.). 150. Aldous S, Gent P, McGeoch G, Nicholson D. The use of troponin in general practice. N Z Med J 2012;125(1357):36-43. 151. Planer D, Leibowitz D, Paltiel O, Boukhobza R, Lotan C, Weiss TA. The diagnostic value of troponin T testing in the community setting. Int J Cardiol 2006 June;107(3):369-75. 152. Tomonaga Y, Gutzwiller F, Luscher TF, Riesen WF, Hug M, Diemand A, et al. Diagnostic accuracy of point-of-care testing for acute coronary syndromes, heart failure and thromboembolic events in primary care: a cluster-randomised controlled trial. BMC Fam Pract 2011 March;12:12. 153. Tanner H, Larsen P, Lever N, Galletly D. Early recognition and early access for acute coronary syndromes in New Zealand: key links in the chain of survival. N Z Med J 2006 Apr;119(1232):(approx 9 p.).

http://www.nice.org.uk/

http://www.journalslibrary.nihr.ac.uk/

http://guidelines.gov/

http://www.sign.ac.uk/index.html

http://www.cochrane.org/handbook

http://www.casp-uk.net/wp-content/uploads/2011/11/CASP-Qualitative-Research-Checklist-31.05.13.pdf

http://www.casp-uk.net/wp-content/uploads/2011/11/CASP-Qualitative-Research-Checklist-31.05.13.pdf

94

154. Law K, Elley R, Tietjens J, Mann S. Troponin testing for chest pain in primary healthcare: a survey of its use by general practitioners in New Zealand. N Z Med J 2006;119(1238):U2082(approx. 8 pages). 155. Sodi R, Hine T, Shenkin A. General practitioner cardiac troponin test requesting: findings from a clinical laboratory audit. Ann Clin Biochem 2007;44(Pt 3):290-3. 156. Tandjung R, Senn O, Rosemann T, Loy M. Diagnosis and management of acute coronary syndrome in an outpatient setting: good guideline adherence in Swiss primary care. J Eval Clin Pract 2013 Oct;19(5):819-24. 157. Goldberg RJ, Steg PG, Sadiq I, Granger CB, Jackson EA, Budaj A, et al. Extent of, and factors associated with, delay to hospital presentation in patients with acute coronary disease (The GRACE registry). Am J Cardiol 2002 Apr 1;89(7). 158. Macdonald SPJ, Nagree Y, Fatovich DM, Phillips M, Brown SGA. Serial multiple biomarkers in the assessment of suspected acute coronary syndrome: multiple infarct markers in chest pain (MIMIC) study. Emerg Med J 2013 Feb;30(2):149-154. 159. Department of Health Western Australia. WA Data Linkage Service[Internet]. Department of Health Government of Western Australia; 2014 [updated 2014 November 17; cited 2014 21 December]. Available from: http://www.datalinkage-wa.org 160. Holman CDJ, Bass AJ, Rouse IL, Hobbs MST. Population-based linkage of health records in Western Australia: development of a health services research linked database. Aust N Z J Public Health 1999 Oct;23(5):453-459. 161. National Centre for Classification in Health. The International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification (ICD-10-AM). 7th ed. Sydney: NCCH; 2010. 162. National Centre for Classification in Health. The International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification (ICD-10-AM). 6th ed. Sydney: NCCH; 2008. 163. National Centre for Classification in Health (NCCH). The Australian Classification of Health Interventions (ACHI) – Seventh Edition - Tabular list of interventions and Alphabetic index of interventions. Sydney: NCCH; 2010. 164. SAS Institute. SAS/STAT user’s guide, version 9.4. Cary, NC: SAS Institute; 2013. 165. Addinsoft. XLSTAT-Pro Version 2014.2. Brooklyn, NY, USA: Addinsoft Inc; 2013. 166. Lloyd-Jones DM, Leip EP, Larson MG, D'Agostino RB, Beiser A, Wilson PWF, et al. Prediction of lifetime risk for cardiovascular disease by risk factor burden at 50 years of age. Circulation 2006 Feb 14;113(6):791-798. 167. Mann S, McBain L. The use of troponin testing in primary care. N Z Med J 2006;125(1357):11-14. 168. Heart Foundation (AU). Warning signs of heart attack. [Internet]. [cited 2014 28 December]. Available from: http://www.heartattackfacts.org.au/warning-signs/ 169. Montalescot G, Dallongeville J, Van Belle E, Rouanet S, Baulac C, Degrandsart A, et al. STEMI and NSTEMI: are they so different? 1 year outcomes in acute myocardial infarction as defined by the ESC/ACC definition (the OPERA registry). Eur Heart J 2007 Jun;28(12):1409-1417. 170. Awad HH, Anderson FA, Jr., Gore JM, Goodman SG, Goldberg RJ. Cardiogenic shock complicating acute coronary syndromes: Insights from the Global Registry of Acute Coronary Events. Am Heart J 2012 Jun;163(6):963-971. 171. Johansson I, Stromberg A, Swahn E. Ambulance use in patients with acute myocardial infarction. J Cardiovasc Nurs 2004;19(1). 172. Raposeiras-Roubin S, Abu-Assi E, Gonzalez-Cambeiro C, Alvarez-Alvarez B, Pereira-Lopez E, Gestal-Romani S, et al. Prognostic influence of prior ischemic heart disease in in-hospital mortality of acute coronary syndromes. Int J Cardiol 2013 Oct 12;168(5):5063-5064.

http://www.datalinkage-wa.org/

http://www.heartattackfacts.org.au/warning-signs/

95

173. Bhatt DL, Eagle KA, Ohman EM, Hirsch AT, Goto S, Mahoney EM, et al. Comparative Determinants of 4-Year Cardiovascular Event Rates in Stable Outpatients at Risk of or With Atherothrombosis. JAMA 2010 Sep 22;304(12):1350-1357. 174. Little P, Dorward M, Warner G, Stephens K, Senior J, Moore M. Importance of patient pressure and perceived pressure and perceived medical need for investigations, referral, and prescribing in primary care: nested observational study. Br Med J 2004 Feb 21;328(7437):444-446A. 175. Walter A, Chew-Graham C, Harrison S. Negotiating refusal in primary care consultations: a qualitative study. Fam Pract 2012 Aug;29(4):488-496. 176. Rolfe A, Burton C. Reassurance After Diagnostic Testing With a Low Pretest Probability of Serious Disease Systematic Review and Meta-analysis. JAMA Intern Med 2013 Mar;173(6):407-416. 177. Sood R, Sood A, Ghosh AK. Non-evidence-based variables affecting physicians' test-ordering tendencies: a systematic review. Neth J Med 2007 May;65(5):167-177. 178. Mann S, Tietjens J, Law K, Elley CR. Troponin testing for chest pain in primary healthcare: A New Zealand audit. N Z Med J 2006;119(1238). 179. Law K, Elley CR, Tietjens J, Mann S. Troponin testing for chest pain in primary healthcare: A survey of its use by general practitioners in New Zealand. N Z Med J 2006;119(1238). 180. Bird S. Acute myocardial infarction: medicolegal issues. Aust Fam Physician 2005 Jun;34(6). 181. Ely JW, Kaldjian LC, D'Alessandro DM. Diagnostic errors in primary care: lessons learned. Journal of the American Board of Family Medicine: JABFM 2012 Jan-Feb;25(1):87-97. 182. Ebell MH. Evaluation of Chest Pain in Primary Care Patients. Am Fam Physician 2011 Mar;83(5):603-605. 183. Buntinx F, Knockaert D, Bruyninckx R, de Blaey N, Aerts M, Knottnerus JA, et al. Chest pain in general practice or in the hospital emergency department: is it the same? Fam Pract 2001 Dec;18(6):586-589. 184. Ruigomez A, Rodriguez LAG, Wallander MA, Johansson S, Jones R. Chest pain in general practice: Incidence, comorbidity and mortality. Fam Pract 2006 April;23(3):167-174. 185. Geelhoed GC, de Klerk NH. Emergency department overcrowding, mortality and the 4-hour rule in Western Australia. Med J Aust 2012 Feb 6;196(2):122-126. 186. Lam Q. Recommendations for communication and flagging of critical laboratory results[Internet]. Australasian Association of Clinical Biochemists; 2014 [updated 2014; cited 2014 December 28]. Available from: http://www.aacb.asn.au/professionaldevelopment/reference-intervals/harmonisation-workshop-2014 187. PathWest Laboratory Medicine WA. Online Test Directory[Internet]. 2014 [updated 2014; cited 2014 December 28]. Available from: http://www.pathwest.com.au/testdirectory/ 188. Grava-Gubins I, Scott S. Effects of various methodologic strategies: Survey response rates among Canadian physicians and physicians-in-training. Can Fam Physician 2008 Oct;54(10):1424-1430. 189. Cho YI, Johnson TP, VanGeest JB. Enhancing Surveys of Health Care Professionals: A Meta-Analysis of Techniques to Improve Response. Eval Health Prof 2013 Sep;36(3):382-407. 190. AIHW Australian GP Statistics and Classification Centre. SAND abstract No. 151 from the BEACH program: Lipid medication use and cardiovascular risk in patients seen in general practice.[Internet]. Sydney: FMRC University of Sydney; 2010 [cited 2014 9 January]. Available from: http://sydney.edu.au/medicine/fmrc/publications/sand-abstracts/index.php

http://www.aacb.asn.au/professionaldevelopment/reference-intervals/harmonisation-workshop-2014

http://www.aacb.asn.au/professionaldevelopment/reference-intervals/harmonisation-workshop-2014

http://www.pathwest.com.au/testdirectory/



96

191. AIHW Australian GP Statistics and Classification Centre. SAND abstract No. 86 from the BEACH program: Diabetes Types 1 and 2 and coronary heart disease[Internet]. Sydney: FMRC University of Sydney; 2006 [cited 2014 9 January]. Available from: http://sydney.edu.au/medicine/fmrc/publications/sand-abstracts/index.php 192. AIHW Australian GP Statistics and Classification Centre. SAND abstract No. 183 from the BEACHprogram: Antiplatelet use and gastrointestinal side effects.[Internet]. Sydney: FMRC University of Sydney; 2012 [cited 2014 9 January]. Available from: http://sydney.edu.au/medicine/fmrc/publications/sand-abstracts/index.php 193. Devlin G, Anderson FA, Heald S, Lopez-Sendon J, Avezum A, Elliott J, et al. Management and outcomes of lower risk patients presenting with acute coronary syndromes in a multinational observational registry. Heart 2005 Nov;91(11):1394-1399. 194. Heeley EL, Peiris DP, Patel AA, Cass A, Weekes A, Morgan C, et al. Cardiovascular risk perception and evidence-practice gaps in Australian general practice (the AusHEART study). Med J Aust 2010 Mar 1;192(5):254-259. 195. Howick J, Cals JWL, Jones C, Price CP, Pluddemann A, Heneghan C, et al. Current and future use of point-of-care tests in primary care: an international survey in Australia, Belgium, The Netherlands, the UK and the USA. BMJ Open 2014;4(8). 196. Shephard M. Point-of-care testing comes of age in Australia. Australian Prescriber 2010 Feb;33(1):6-9. 197. Kramer CM. Avoiding the Imminent Plague of Troponinitis The Need for Reference Limits for High-Sensitivity Cardiac Troponin T. J Am Coll Cardiol 2014 Apr 15;63(14):1449-1450. 198. Galbraith PD, T Larsen E, Maxwell CJ, Schopflocher DP, Svenson LW, Ghali WA. "Troponinitis" without an accompanying myocardial infarction code. Circulation 2005 May 24;111(20):E340-E340. 199. Yip TPY, Pascoe HM, Lane SE. Impact of high-sensitivity cardiac troponin I assays on patients presenting to an emergency department with suspected acute coronary syndrome. Med J Aust 2014 Aug 4;201(3):158-161. 200. Bonevski B, Magin P, Horton G, Foster M, Girgis A. Response rates in GP surveys: Trialling two recruitment strategies. Aust Fam Physician 2011 Jun;40(6):427-430. 201. National Prescribing Service. MedicineInsight. 2013 [updated 13 Feb 2013; cited 2015 24 Jan]. Available from: http://www.nps.org.au/about-us/what-we-do/medicineinsight 202. Clinical Practice Research Datalink Group. Observational Data[Internet]. London UK: Medicines and Healthcare Products Regulatory Agency; 2014 [cited 2014 October 11]. Available from: http://www.cprd.com/ObservationalData/ 203. Britt H MG, Henderson J, Bayram C, Valenti L, Harrison C. General practice activity in Australia 2011-12. General practice series no.31[Internet]. Sydney: Sydney University Press; 2013 [cited 2014 October 11]. Available from: http://hdl.handle.net/2123/8675 204. Shaw N, Hegedus G. The national programme for information technology - The GP as gatekeeper - a bastion worth fighting for? Br J Gen Pract 2005 Feb;55(511):85-86. 205. Rask KJ, Deaton C, Culler SD, Kohler SA, Morris DC, Alexander WA, et al. The effect of primary care gatekeepers on the management of patients with chest pain. Am J Manag Care 1999 Oct;5(10):1274-1282. 206. Emery J, Chiang P. The role of risk tools in diagnosing cancer in primary care. Aust Fam Physician 2014 Aug;43(8):508-12.




http://www.nps.org.au/about-us/what-we-do/medicineinsight

http://www.nps.org.au/about-us/what-we-do/medicineinsight

http://www.cprd.com/ObservationalData/

http://hdl.handle.net/2123/8675

97

APPENDICES

Appendix 1: Search strategy

1.1 Medline, EBM Review and Embase combined search strategy

1. exp Cardiovascular diseases/ (4224045) 2. exp Acute Coronary Syndrome/ (34096) 3. exp Angina, Unstable/ or exp Myocardial Infarction/ or exp Coronary Disease/ or exp Acute Coronary Syndrome/ (684955) 4. exp Coronary Artery Disease/ (228567) 5. exp Coronary Thrombosis/ (11895) 6. exp Dyspnea/ (80823) 7. exp Arrhythmia/ (444708) 8. exp Fatigue/ (142053) 9. exp syncope/ (15776) 10. 8 and heart.mp. [mp=title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept, rare disease supplementary concept, unique identifier] (20123) 11. exp Chest Pain/ (102494) 12. general practitioner.mp. or exp General Practitioners/ (80222) 13. family practice.mp. or exp Family Practice/ (121024) 14. primary health care.mp. or exp Primary Health Care/ (191722) 15. exp Physicians, Primary Care/ or exp Physicians, Family/ (72149) 16. 12 or 13 or 14 or 15 or primary care.mp (411562) 17. troponin.mp. or exp Troponin C/ or exp Troponin/ or exp Troponin T/ or exp Troponin I/ (47191) 18. 16 and 17 (159) 19. delay.mp. or exp delay/ (252335) 20. Time Factors.mp (1029166) 21. misdiagnosis.mp. or Diagnostic Errors/ (77207) 22. Resuscitation/ or Emergency Medical Services/ or Time Factors/ or prehospital.mp. (1462509) 23. Heart Arrest/di [Diagnosis] (2236) 24. prehospital care.mp. (2496) 25. Point-of-Care Systems/og, st, td, ut [Organization & Administration, Standards, Trends, Utilization] (1322) 26. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 9 or 10 or 11 (4380798) 27. 22 or 23 or 24 (1464347) 28. 19 or 20 or 21 (1612065) 29. 16 and 17 and 26 (150) 30. 16 and 17 and 26 and 28 (10) 31. 17 and 26 and 28 (1682) 32. 16 and 17 and 28 (11) 33. 16 and 26 and 28 (1756) 34. 17 and 27 and 28 (1690) 35. 16 and 17 and 25 (1) 36. 16 and 25 and 27 (4) 37. 16 and 17 and 27 (24)

98

38. 17 and 25 and 27 (9) 39. 16 and 26 and 27 and 28 (433) 40. 18 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 (3915) 41. limit 40 to (humans and yr="1995 -Current") (3223) 42. remove duplicates from 41 (2709) 1.2 Web Of Science search strategy

#1 Topic=(general pract* or family pract* or primary care or family physician) (282671) #2 ts =(acute coronary syndrome) OR ts=(chest pain) or ts=(cardiovascular disease) or ts=(coronary artery disease) or ts=(coronary heart disease) or ts=(Angina) or ts=(Myocardial Infarction) (466241) #3 ts =(Coronary Thrombosis) or ts=(Dyspnea) or ts=(arrhythmia) or ts=(Fatigue) or ts=(syncope) (224322) #4 #3 OR #2 (659247) #5 ts =(troponin) (20445) #6 ts =(delay) or ts =(error*) (1083775) #7 ts =(Point-of-Care Systems) (2582) #8 ts =(prehospital care) (3077) #9 #5 and #2 and #1 (187) #10 #9 and #6 (11) #11 #6 and #5 and #2 (246) #12 #6 and #5 and #1 (13) #13 #8 and #6 and #5 (7) #14 #7 and #5 and #1 (4) #15 #8 and #7 and #1 (3) #16 #8 and #5 and #1 (8) #17 #8 and #7 and #5 (5) #18 #7 and #6 and #5 and #1 (2) #19 #5 and #1 (243) #20 #6 and #5 and #2 and #1 (11) #21 #8 and #6 and #2 and #1 (118) #22 #6 and #2 and #1 (694) #23 #22 or #21 or #20 or #19 or #18 or #17 or #16 or #15 or #14 or #13 or #12 or #11 or #10 or #9 (1163) #24 topic=(poct) (573) #25 #24 or #7 (2976) #26 #25 and #2 and #1 (35) #27 #23 or #26 (1167)

.2.3 Scopus search strategy

1. (TITLE-ABS-KEY(Cardiovascular diseases) OR TITLE-ABS-KEY(Acute Coronary Syndrome) OR TITLE-ABS-KEY(Angina,Unstable) OR TITLE-ABS-KEY(Myocardial Infarction) OR TITLE-ABS-KEY(Coronary Disease) OR TITLE-ABS-KEY(Coronary Artery Disease) OR TITLE-ABS-KEY(Coronary Thrombosis) OR TITLE-ABS-KEY(Dyspnea) OR TITLE-ABS-KEY(chest pain) OR TITLE-ABS-KEY(Arrhythmia) OR TITLE-ABS-KEY(Fatigue) OR TITLE-ABS-KEY(syncope/)) AND PUBYEAR > 1989 (985,377)

http://www.scopus.com/search/history/results.url?origin=searchhistory&shid=4

99

2. (TITLE-ABS-KEY(General practice) OR TITLE-ABS-KEY(Physicians,Family) OR TITLE-ABS-KEY(Physicians,Primary Care) OR TITLE-ABS-KEY(Primary Health Care) OR TITLE-ABS-KEY(General Practitioners)) (430,531) 3. TITLE-ABS-KEY(troponin) AND PUBYEAR > 1989 (24,847) 4. (TITLE-ABS-KEY(delay) OR TITLE-ABS-KEY(misdiagnosis) OR TITLE-ABS-KEY(Diagnostic Errors)) (520,036) 5. (TITLE-ABS-KEY(Resuscitation) OR TITLE-ABS-KEY(prehospital) OR TITLE-ABS-KEY(Time Factors) OR TITLE-ABS-KEY(Emergency Medical Services)) (1,900,643) 6. 1 and 2 and 3 (209) 7. 1 and 2 and 3 and 4 (15) 8. 1 and 3 and 4 (328) 9. 2 and 3 and 4 (18) 10. 3 and 4 and 5 (111) 11. TITLE-ABS-KEY(point of care testing) (9,804) 12. 2 and 5 and 11 (167) 13. 2 and 3 and 5 (67) 14. 3 and 5 and 11 (77) 15. 2 and 3 and 4 and 5 (9) 16. 2 and 3 (253) 17. 1 and 2 and 4 and 5 (404) 18. 6 or 7 or 8 or 9 or 10 or 12 or 13 or 14 or 15 or 16 or 17 (1081)





100

Appendix 2: Additional literature sources

2.1 Grey literature sources

SIGLE (System for Information on Grey Literature) database. Available at www.opensigle.inist.fr NTIS (National Technical Information Service). Available at www.ntis.gov Proquest Dissertations and Theses BIOSIS Previews

2.2 Journals and conference proceedings handsearched:

2.1.1 Australian primary care, emergency medicine and cardiology journals

1. Emergency Medicine Australasia

2. Medical Journal of Australia

3. Australian Family Physician

4. New Zealand Medical Journal

2.2.2 International primary care journals (chosen according to impact factor)

1. Annals of Family Medicine

2. Scandinavian Journal of Family Practice

3. Family Practice

4. Journal of the American Board of Family Medicine

5. Biomed Central Family Practice

6. British Journal of General Practice

7. Quality and safety journals

8. BMJ Quality and Safety

9. Clinical risk

2.2.3 Other journals

1. Journal of Clinical Pathology

http://www.opensigle.inist.fr/

http://www.ntis.gov/

101

Appendix 3: GP cohort questionnaire

INFORMATION SHEET AND CONSENT FORM Troponin testing for diagnosis of acute coronary syndromes in primary care

Dear Colleague, Last week you were sent an information sheet and a questionnaire relating to troponin testing ordered by you. This test was requested on (date) on your patient (name). The study is being conducted by the School of Primary, Aboriginal and Rural Health Care (SPARHC) at the University of Western Australia in collaboration with the laboratories of St John of God Pathology and Pathwest. We are analysing all troponin test results ordered by general practitioners in the six month period starting from March 2009. We are collecting information on the clinical characteristics of these patients, their cardiovascular risk factors, the GP’s perception of the cause of their symptoms and their management after the test result was known. From this we hope to understand if troponin testing is safe to be done in general practice. We hope to show that: • Troponin testing can be safely used in general practice on a certain group of patients, who are likely to be those who present late in the course of their illness or with atypical symptoms; • GPs have good knowledge of the use and limitations on troponin testing. We would be very grateful if you could complete this survey. We estimate that it will take less than 5 minutes to complete including time to access records. Thank you very much for your assistance. Dr Helen Wilcox Professor Jon Emery Senior Lecturer, Discipline of General Practice Head of School School of Primary, Aboriginal and Rural Health Care Faculty of Medicine, Dentistry & Health Science University of Western Australia 328 Stirling Highway, Claremont, WA 6010 Telephone 08 9449 5121 Facsimile 08 93846238 [email protected]

mailto:[email protected]

102

Consent form

I, (Name of GP), have read the information provided and any questions I have asked have been answered to my satisfaction. I agree to participate in this activity, realising that I may withdraw at any time without reason and without prejudice. I understand that all information provided is treated as strictly confidential and will not be released by the investigator. The only exception to this principle of confidentiality is if documents are required by law. I have been advised as to what data is being collected, what the purpose is, and what will be done with the data upon completion of the research.

I agree that research data gathered for the study may be published provided my name or other identifying information is not used.

______________ __________________ Participant Date

PLEASE FAX THIS CONSENT FORM WITH COMPLETED QUESTIONNAIRE TO:………….. The Human Research Ethics Committee at the University of Western Australia requires that all participants are informed that, if they have any complaint regarding the manner, in which a research project is conducted, it may be given to the researcher or, alternatively to the Secretary, Human Research Ethics Committee, Registrar’s Office, University of Western Australia, 35 Stirling Highway, Crawley, WA 6009 (telephone number 6488-3703). All study participants will be provided with a copy of the Information Sheet and Consent Form for their personal records. This information will be sent to a researcher who is employed within (laboratory). In order to maintain your privacy and that of your patient, the researcher will remove all identifying information prior to forwarding test requests and results to the Chief Investigator who will analyse the data. The patients on whom the test was performed will not be made aware of their, or your, involvement in the study. Data will be kept within secure storage facilities at the laboratory and at the School of Primary, Aboriginal and Rural Health Care and access to this is only available to the study investigators. The data will be destroyed after the completion of the study which is estimated to take two years. Information will not be released to any third party unless required by law. Approval has been granted by the University of Western Australia’s Human Ethics Research Committee for this study. The study conforms to the NHMRC National Statement on Ethical Conduct in Human Research (2007). Funding for the project has been granted through the Primary Health Care Research, Evaluation and Development which is supported by the Commonwealth Government. Neither of the participating laboratories have any financial interest in the project. Voluntary participation and withdrawal from this project You are free at any time to withdraw consent to further participation without prejudice in any way. You need give no reason for your decision and the record of your participation will be destroyed on your request.